Sie sind auf Seite 1von 52

advantage

TM

excellence in engineering simulation


volume V issue 1 2011

Sustainable
Design

INNOVATE REDUCE SAVE


QUICKLY PROTOTYPES energy
page 6 page 12 page 17
2X THE DESIGN CANDIDATES.
NO TRADE OFFS.
ANSYS MECHANICAL
WITH NVIDIA® TESLA™ GPUs

Accelerate your design with ANSYS Mechanical 13.0 and NVIDIA® Tesla™ GPUs (graphics
processing units). The GPU-accelerated ANSYS structural mechanics solution enables you
to double your simulation speed, while reducing prototyping time and shortening design
cycles—doubling your design candidates within the same design cycle with no trade offs.

This acceleration is possible when you run ANSYS Mechanical on a workstation equipped
with a Tesla GPU featuring many “must have” capabilities for technical computing.
These Tesla GPU-powered workstations deliver performance increases that accelerate the
performance of computing with CPUs alone.

To learn more about ANSYS acceleration on Tesla GPUs, visit www.nvidia.com/tesla.


©2011 NVIDIA Corporation. NVIDIA, the NVIDIA logo, and Tesla are trademarks or registered trademarks of NVIDIA Corporation in the United States and
other countries. Other company and product names may be trademarks of the respective companies with which they are associated. All rights reserved.
TABLE OF CONTENTS

Table of Contents
Sustainable design
4 Best Practices
Green Design without Compromise
Companies can gain a sustainable competitive advantage and be
environmental stewards — through robust design and optimization.

6 transportation
4
Cleaner, Greener Engine Design
Cummins uses simulation to reduce weight, improve fuel
economy and reduce emissions of engines.

9 Construction
Brussels’tainable
Simulation helps to determine potential environmental impact
in preparation for a massive renovation in Brussels.

12 ENERGY
Catch the Next Wave
Hydrodynamic simulation helps to deliver two- to three-times wave
power efficiency improvement.
6
15 Energy
Reacting to Emissions
Fluids simulation helps to speed up research into chemical-looping
combustion capable of reducing fossil fuel emissions.

17 construction
Meeting Green Building Design Goals with
Engineering Simulation
Simulation is driving innovation in HVAC design for an assembly hall in Saudi Arabia. 9

19 Offshore
Don’t Rock the Float! 12
Fluid–structure interaction allows designers to assess impact of waves
on freshwater and offshore systems.

22 Energy: Wind
Energizing the Wind Industry
Increased complexities require a system-level approach
in designing and evaluating wind turbines.

24 Energy: WIND
More Power to You
Simulation helps Indar to design one of the world’s highest-efficiency
permanent magnet wind turbine generators. 17

24

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 1


TABLE OF CONTENTS

27 Energy: Wind
Where the Wind Blows
Engineering simulation plays a role in getting the most power
from wind farms by predicting the best available locations.

30 ENERGY: Wind
Second Wind
Advanced turbulence models lead to optimized wind turbine spacing. 30

Simulation@Work
32 Aerospace
Fast Lane to Sky High
Fluid flow simulation software co-pilots design of
production prototype roadable aircraft.

35 electronics 32
Successful Launch
Circuit and field tools combine to optimize satellite multiplexer design,
reducing time from 10 weeks to two days.

38 Scientific Equipment
Glass Jaw
Simulation helps to solve collimator jaw design problem in 38
the Large Hadron Collider.

Departments
41 ANALYSIS TOOLS
Designing Batteries for Electric Vehicles
Numerical simulation can be used to accelerate battery development
and address safety concerns.

44 Academic
Microbubbles Keep Green Energy Blooming 41
Algae-derived biofuel production gets a strong pulse from flow simulation.

46 Academic
Reforming a Fuel Cell Modeling Process
Coupling flow simulation with complex chemistry tools brings a united
front to analyzing leading-edge energy systems.

48 Tips and Tricks


Accelerating CFD Solutions
Several recent enhancements in ANSYS FLUENT solver
capabilities accelerate convergence and reduce solution time.

46

48

2 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


TABLE OF CONTENTS

In Memoriam For ANSYS, Inc. sales information,


call 1.866.267.9724
John Krouse, senior editor and industry analyst Email the editorial staff at
ansys-advantage@ansys.com.
for ANSYS Advantage, passed away on
November 6, 2010. He was editorial director and Executive Editor
Fran Hensler
one of the originators of ANSYS Solutions, the
Managing Editor
corporate publication that preceded ANSYS Chris Reeves
Advantage. His credits included being editor and Art Director
publisher with Penton Media for 18 years, Dan Hart

John Krouse where he contributed to such publications Editors


Erik Ferguson
as Machine Design and Computer Aided Shane Moeykens
Mark Ravenstahl
Engineering. Krouse founded the communications consulting firm
Krouse Associates in 1994. His understanding of CAD/CAM/CAE was Ad Sales Manager
Helen Renshaw
extensive, and he authored many books and articles on the topic.
Editorial Contributor
ANSYS Advantage readers and collaborators alike will miss ANSYS North America
Support and Services
his ability to explain complex engineering concepts in an
entertaining manner, as well as the keen perception he brought to Editorial Advisor
Tom Smithyman
light in his editorials.
Designer
Miller Creative Group

Circulation Manager
Sharon Everts

About the Cover


Turbulence can greatly affect the
ANSYS Webinar Series placement of wind turbines on a farm.
This and other ANSYS Advantage
features focus on solutions for
sustainable design.
Learn how software from ANSYS
• streamlines your workflow Simulation image courtesy Fluid and
Energy Engineering GmbH & Co. KG.
• is applied to a wide range of applications Photo © iStockphoto.com/Mienny.
• is being enhanced to help you develop better
products and processes
Neither ANSYS, Inc. nor the senior editor nor Miller Creative
Group guarantees or warrants accuracy or completeness of
Every week ANSYS holds several online seminars to provide the material contained in this publication.
information on our product capabilities, demonstrate ANSYS, ANSYS Workbench, Ansoft Designer, CFX,
AUTODYN, FLUENT, GAMBIT, POLYFLOW, Airpak,
the use of our software and display a broad range of DesignSpace, FIDAP, Flotran, Iceboard, Icechip, Icemax,
Icepak, FloWizard, FLOWLAB, G/Turbo, MixSim, Nexxim, Q3D
industry applications. Extractor, Maxwell, Simplorer, Mechanical, Professional,
Structural, DesignModeler, TGrid, AI*Environment, ASAS,
AQWA, AutoReaGas, Blademodeler, DesignXplorer, Drop
Register for upcoming events and view recorded webinars: Test, ED, Engineering Knowledge Manager, EKM, Emag,
www.ansys.com/webinars Fatigue, Icepro, Icewave, Mesh Morpher, ParaMesh, TAS,
TASSTRESS, TASFET, TurboGrid, Vista, VT Accelerator, CADOE,
CoolSim, SIwave, Turbo Package Analyzer, RMxprt, PExprt,
HFSS, Full-Wave SPICE, VerifEye, QuickEye, Optimetrics, TPA,
AnsoftLinks, ePhysics, Simulation Driven Product Development,
Smart Engineering Simulation and any and all ANSYS, Inc.
brand, product, service, and feature names, logos and slogans
are registered trademarks or trademarks of ANSYS, Inc. or its
subsidiaries located in the United States or other countries.
ICEM CFD is a trademark licensed by ANSYS, Inc. All other
brand, product, service and feature names or trademarks are
the property of their respective owners.

ANSYS, Inc.
Southpointe
275 Technology Drive
Canonsburg, PA 15317
U.S.A.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 3


Best Practices

Green Design
without Compromise
Companies can gain a sustainable competitive advantage and be
environmental stewards — through robust design and optimization.
By Thierry Marchal, Industry Director, ANSYS, Inc.

The Need for Sustainability Reluctance to Sacrifice Current Standards


Throughout much of the world, people expect that life Without a doubt, cars could be more environmentally
will become better, easier and more comfortable for friendly, but if you surveyed the planet, would you find
succeeding generations. Over the past century, privileges people willing to pay more to achieve this objective? Most
once limited to the wealthy have become commonplace; would be reluctant to accept significant reductions in
indeed, many are now considered a necessity. Owning a performance, speed or acceleration. Green building is a
car, flying around the world, accessing a computer, common aspiration, but will the occupants tolerate being
connecting to the internet from anywhere, and using a cell a bit less warm in winter, and a bit more warm in
phone are routine activities, even in emerging countries. summer? Renewable energy is a “must do” only if the
However, the world can’t sustain such exponential resulting power is less expensive and at least as
growth forever. The planet has received a wake-up call stable. An admirable goal: Can we reach it without
about addressing overconsumption. Many scientists personal compromise?
state emphatically that we are putting our planet, and the Common sense dictates that we can’t meet such
human species, in danger. expectations using traditional approaches and tech-
The population is rapidly draining the world’s fossil nologies. Real breakthroughs may be necessary to
energy sources; at the same time, emissions from these achieve sustainability goals in these areas:
fuels are negatively impacting the environment and its • Better energy management: Fossil fuels remain
natural evolution. Huge amounts of waste material are an important energy source, but they could be
being stored with the hope that the earth can cope with it. used more effectively and selectively. Sources
Most people are aware that the globe cannot sustain of renewable energies should be exploited.
these practices forever. While a growing number of people Transportation, construction and manufacturing
are urging the world’s population to do something must become more energy efficient.
about it, many individuals are unwilling to sacrifice their • Pollution reduction: Twenty-first century lifestyles
comfortable lifestyle. result in the release of huge quantities of various
materials into the environment, making absorption
and recycling of these substances difficult without a
direct impact on the planet.
• Product lifecycle: After a lifetime of use, products
quickly become waste material. Shortened product
life increases the amount of waste. Processing
recycled materials into new products in a cost-
effective manner can benefit greatly from modern
simulation technology.

Virtual Experimentation, Virtual Testing, Virtual Prototyping


If sustainable products and processes today are
merely “nice to have,” it is likely green design will become
mandatory in the future, legislated and forced by market
pressures to develop better solutions. Pioneers and
Engineering for sustainability includes designing more efficient ways
to treat waste, such as studying settling in a wastewater treatment tank.
leading companies are already preparing for this evolution
Courtesy MMI Engineeering. by designing sustainable solutions. But the challenge is

4 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Best Practices

Background photo © iStockphoto.com/ooyoo


Engineering simulation can help organizations meet regulations and standards. This study was used to assess the impact of fume cupboard
discharges on surrounding buildings and the environment.
Courtesy BDP Engineering.

daunting. Environmentally friendly products may require standard technology. Designing smart, green products to
major innovations that could impact both cost and robust- a high level of product integrity requires investigating a
ness. Innovation usually requires intense experimentation growing number of parameters — such as dimension,
and optimization. Real breakthrough solutions need materials and operating conditions — to provide
systematic testing to ensure that new designs behave designers with enough freedom to meet numerous
properly throughout the entire lifecycle. Experimentation constraints. Robust design includes identification of
and testing processes are typically time intensive, costly influential parameters and evaluation of design sensitivity
and difficult to manage in a highly competitive world. based on performing variations.
Best-in-class companies are addressing this issue by The combination of parametric studies and sensitivity
switching to the virtual world. analysis for a large number of parameters using advanced
Numerous reports and industry studies, some found tools (robust design and optimization) opens the door to a
in this issue of ANSYS Advantage, suggest that an order new era of design.
of magnitude more experimentation can be done virtually
for a smaller cost. Engineering teams can vary numerous Engineering Simulation: A Profitable Green Technology
parameters to identify the best combination of sustainable Engineering simulation has proven its cost-
design, minimal cost and maximal performance. Leading effectiveness in developing innovative products. This
companies are systematically testing same technology can be used
their virtual product prototypes against to design greener products
conditions that would be experienced and processes. Pioneering com-
10
throughout the lifecycle to ensure that panies are already creating a
product behavior will satisfy end users. cleaner and more profitable
P12 - FKalotte [N]

9
Continually increasing computational future by changing their design
power combined with high-efficiency 8
processes and widely adopting
solvers can contribute to the effort: 7
simulation to transform the
They might enable the number or green challenge into a major
complexity of virtual tests to double 6 business opportunity. This issue
0.35
every 18 months without impact on of ANSYS Advantage illustrates
cost or time to market. P1 0.4
0- 0.55 h o w c o m p a n i e s a re u s i n g
pa 0.65
r_R
d 0.45
0.75
0.85 P6 - p a r_ R k comprehensive integrated multi-
Robust Design and Optimization physics from ANSYS to maintain
Designing smart green products to a high level
Sustainability adds a new of product integrity requires advanced tools that
corporate responsibility while
dimension that is difficult to handle with investigate a growing number of parameters. increasing profit. n

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 5


transportation
Image © iStockphoto.com/PeskyMonkey

Cummins ISX15
heavy-duty engine

Cummins ISF3.8 light


commercial vehicle engine

Cleaner, Greener
Engine Design
Cummins uses simulation to reduce weight,
improve fuel economy and reduce emissions of engines.
By Bob Tickel, Director of Structural and Dynamic Analysis, Cummins Inc., Columbus, U.S.A.

The phrase “environmentally responsible” doesn’t seem thing to do. The company wants the trucking industry to
to fit into a sentence with the terms 18-wheel tractor trailer be viewed as an environmental steward and champion,
and heavy-duty truck. As a world-leading manufacturer of not one of the “bad guys.”
commercial engines and related systems, Cummins Inc. is
working to change that perception one design at a time, Building the Truck of the Future
developing next-generation technologies that are revolu- In recognition of its environmental technology
tionizing the international trucking industry. leadership, Cummins recently received nearly $54 million
Using software from ANSYS, Cummins is developing in funding from the U.S. Department of Energy (DOE) to
and testing radical improvements in engine design, support two projects aimed at improving fuel efficiency in
including the use of alternative materials and smaller both heavy- and light-duty vehicles.
engine footprints that reduce weight, improve fuel About $39 million will fund Cummins’ development of
economy and reduce emissions — while also boosting a new “supertruck” — a highly efficient and clean diesel-
performance. The work of the corporate research and fueled Class 8 (heavy-duty) truck that is expected to set a
technology organization focuses on developing new, new industry standard for green technology. Another
environmentally responsible technologies for the $15 million in funding will support the development of
company’s core engine business. advanced-technology powertrains for light-duty vehicles.
Cummins’ recent product development efforts target a The resulting improvements in engine system efficiency
great deal of attention on fuel economy and emissions — will mean significantly lower fuel and petroleum
and with good reason. Government environmental consumption by these vehicles as well as a significant
standards grow more challenging every year. Because reduction in greenhouse gas emissions.
commercial trucking is a low-margin business, every Though details of this work are proprietary, Cummins
improvement that Cummins makes in fuel economy adds is using software from ANSYS to simulate engine
to its customers’ successes. Beyond these practical performance and achieve the ambitious goals defined by
considerations, Cummins invests in environmentally the DOE: Improve Class 8 vehicle freight efficiency by
responsible engine technologies because it is the right 50 percent and achieve a 40 percent improvement in fuel
economy in light-duty vehicles.

6 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


transportation

Selective catalytic
reduction (SCR) catalyst Decomposition filter

Cummins particulate filter

Diesel exhaust fluid (DEF)


dosing valve

Cummins new aftertreatment system


aimed at meeting emissions standards

To illustrate how Cummins leverages ANSYS software Using Simulation to Rev Up Ongoing Engine Improvements
on a daily basis, consider this recent product introduction: Environmental standards and customer needs are an
the next-generation ISX15 engine design, developed ever-evolving target. So the Cummins team uses
specifically to meet stringent new emissions standards engineering simulation to improve engine features and
from the U.S. Environmental Protection Agency (EPA). boost performance. Design changes that reduce
These regulations, which call for near-zero emission emissions and fuel consumption often result in higher
levels of nitrogen oxide (NOx) and particulate matter (PM), temperatures and pressures within the engine, which
place new demands on truck manufacturers and require Cummins engineers to continually test the limits of
fleet managers. conventional engine designs.
Cummins used ANSYS technology to develop or Most materials used in diesel engines exhibit reduced
enhance a number of features in the ISX15 design, strength as temperature rises. The combination of high
including the new Cummins aftertreatment system that pressure/temperature and reduced strength places stress
incorporates a revolutionary diesel particulate filter (DPF) on components such as the cylinder head, a geometrically
targeted at meeting the new emissions standards. complex casting that serves many functions including
The DPF removes diesel particulate matter or soot transferring engine oil and coolant, intaking air and
from the exhaust gas of a diesel engine. Cummins used exhausting gas. The head also houses the fuel injector
ANSYS software to simulate typical operating loads and and valve train components; it must contain hot combus-
make predictions about the new DPF’s performance and tion gases during the cylinder firing event. To address the
reliability. Engineering simulation predicted both peak multiple forces that come into play, Cummins engineers
temperatures and temperature distribution inside this use multiphysics software from ANSYS to combine
component under a range of operating conditions. These thermal and structural analyses in their work on this
thermal analyses were critical as they revealed the peak sophisticated engine component.
temperatures and temperature gradients within the filter, Simulation tools from ANSYS enable the Cummins
which ultimately determine the thermal fatigue and life of engineering team to evaluate the use of new materials
the component. across the entire topological surface of cylinder heads
Varying temperatures within the DPF result in thermal and other engine components. ANSYS allows the
stresses; if this component is not designed properly, tem- Cummins team to quickly and easily answer questions
perature variations can lead to component failure. While it
might take months or years for failure to occur in real-
world use, software from ANSYS enabled Cummins to
quickly simulate the effects of years of field usage and
make predictions about how the DPF would hold up over
time. This gave the engineering team a high degree of
confidence as they designed the component and installed
it into customer vehicles.
Subsequent field and bench tests confirmed the simu-
lation predictions, as did actual road results. Engineers
can rarely rely on virtual testing alone, but software from
ANSYS confirmed that the Cummins design process was By using software from ANSYS to predict maximum temperatures
and temperature distribution within its new diesel particulate filter
moving in the right direction — and that the results would under diverse conditions, Cummins engineers can ensure that the
be as team members expected. manufactured component will have acceptable durability.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 7


transportation

such as, “What happens if we manufacture the cylinder


h e a d o u t o f t h i s n e w m a t e r i a l ? W h a t a re t h e
performance gains?” without the need to actually machine
and test new components. Working in a simulated
environment gives these engineers the freedom to tweak
existing engine designs — as well as to arrive at some
“clean-sheet” designs that may have the power to revolu-
tionize engine performance.
The team recently used software from ANSYS to
predict temperature profiles in cylinder heads under
various operating conditions. The thermal results were
Cummins engineers can easily test the effectiveness of new materials,
used to determine how this component would perform in new designs and other innovations — and predict their long-term effect
the real world, as well as its threshold for both low-cycle on overall engine performance. For example, Cummins engineers used
software from ANSYS to simulate the impact of new materials on a
and high-cycle fatigue. cylinder head design.
Since cylinder heads are long-lead-time, expensive
components, the Cummins team must be sure that a new
design is right before moving forward. Using simulation have to rely instead on very expensive, time-consuming
tools from ANSYS, Cummins engineers can not only endurance tests to verify the engineering team’s designs.
create more exacting test results but also pursue ANSYS is a key enabler to Cummins, reducing the
increased productivity by incorporating capabilities such company’s overall cost of development by minimizing its
as conjugate heat transfer modeling in their designs. investment in physical engine testing. This approach
Without analysis-led design to introduce new allows the design of a more environmentally friendly
materials and new part configurations, Cummins would engine without compromising cost or performance. n

Shifting Gears: ANSYS Creates a Cultural Change


With so many engine components and performance to accommodate our needs — as well as the multi-
aspects to consider, the Cummins engineering team physics capabilities to link the results of our various
must perform a wide range of structural simulation simulations.”
and analysis. According to Bob Tickel, director of Tickel noted that the efficiency and cost effective-
structural and dynamic analysis at Cummins, ness of engineering simulation has resulted in a
software from ANSYS is the only single-vendor tool complete cultural change at Cummins. “The ease of
with the technical breadth and depth to meet using simulation tools from ANSYS has helped to
this challenge. transform our organization from a test-centric culture
Recently, the Cummins corporate research to an analysis-centric one,” said Tickel.
and technology team began the switch from the “When investigating a new material or other design
traditional interface for ANSYS software to the enhancement, traditionally we would build new parts
ANSYS Workbench platform, a decision based and conduct physical tests as a first step — which
primarily on the product’s improved geometry import, represented a time- and cost-intensive approach,”
cleanup and meshing capabilities. Tickel anticipates he said. “Today, we focus more attention on upfront
a 50 percent reduction in throughput time based analysis, only moving to part-building and testing for
upon the move to ANSYS Workbench. those design improvements that we can verify first using
“The ANSYS Workbench environment provides ANSYS tools. This new cultural approach has not only
access to the best multiphysics tools we need to saved us time and money, but allowed us to selectively
conduct many types of simulation and analysis,” said focus our attention on those design enhancements
Tickel. “Whether our need is thermal, structural, that are shown to hold the greatest promise for
dynamic or static engineering analysis, ANSYS revolutionizing future engine designs.”
Workbench provides the flexibility and versatility

8 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Construction

Brussels’tainable
The State Administrative City in
Brussels was completed in 1983
and is now largely unused.
Photo courtesy Michael Uyttersprot.

Simulation helps to determine potential environmental impact


in preparation for a massive renovation in Brussels.
By Tin Meylemans, Architectural Engineer, and Jean-Pierre Demeure,
Head of Urban Planning, City of Brussels, Belgium
Arnaud Boland and Quentin Hamoir, Support Engineers, ANSYS, Inc.

In 1992 during the Earth Summit in Rio de Janeiro,


200 countries — including Belgium — adopted Agenda
21, an action program for the 21st century to enable
globally sustainable development on the planet. Among
the program’s main objectives is responsible management
of natural resources and harmonious urban development.
As the federal capital of Belgium and an administrative
center of the European Union, the City of Brussels has
been a leader by example. It is also a prime contributor to
generating awareness of how large cities can naturally
evolve to environmentally friendly urban growth.
G u i d e d b y A g e n d a 2 1 a n d t a rg e t i n g u r b a n
revitalization, the city undertook renovation of the State
Administrative City, or CAE (Cité Administrative de l'Etat),
an important site in downtown Brussels. Often recognized
as the finest architectural achievement in Brussels since
World War II, CAE is nevertheless an urban planning
defeat, as it eliminated an entire district and broke the
urban fabric with its 140,000 square meters of office
space. As the Belgian federal government evolved, the
administrative city has not been used since the early
2000s, leaving this huge site unoccupied. Its renovation
had become a city priority. Today the site is undergoing a
massive renovation project that combines housing,
offices, shops, parks and a school.
Before any major project is constructed, Belgian law
requires an environmental impact study that includes
economic and social factors, environmental assessments,
mobility studies, health influences and noise level effects. Initial (top) and modified (bottom) geometries of Brussels’ CAE

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 9


Construction

Initial design Recommended revision


Engineering simulation was performed to determine pedestrian comfort levels (with yellow and red indicating discomfort).

The City of Brussels selected the ANSYS office in Belgium Brussels in which the building must be situated to satisfy
to investigate wind flow patterns for the new site and local planning regulations). A second configuration, called
compare different options for building locations — via the modified geometry, was studied to quantify the impact
engineering simulation. of architectural changes, such as towers being built on
Wind analysis is useful in studying many parameters stilts, or architectural details, such as carvings or
related to a site’s user comfort, in this case including openings placed on different levels/stories of the
high-rise buildings and green spaces. In some instances, structure. Because plans for reuse of the area were not
the building acts as a screen and improves comfort, but in complete at the time of analysis, this study would show
others the wind pattern could cause significant discomfort where changes could have an impact.
— and might even be dangerous — to pedestrians and
occupants. Several important phenomena can occur, such
as channeling and wedge effects, that lead to local wind
acceleration and increased turbulence. These phenomena
can contribute to discomfort for people located at ground
level, terraces and balconies. The wind effects can
also damage vegetation and generate unwanted
pressure effects on buildings that cause whistling or
material damage.
A typical comfort criteria from the TNO (an
independent research organization in the Netherlands
whose aim is to apply scientific knowledge to strengthen
the innovative power of industry and government) requires
that the wind speed in an area 1.75 meters above the
ground should not exceed 5 meters/second for more than
220 hours per year. In identifying areas of discomfort, the
standard must be expressed in terms of amplification
factor (AF), the ratio between velocity without and with the
building. A factor greater than 1 means the flow is
accelerated by the buildings. Considering Brussels’
meteorological data over the last 10 years, an AF of less
than 0.88 satisfies the most-restrictive comfort criterion.
Project engineers investigated two configurations for
environmental impact along with a possible need to
minimize the negative impacts of wind. The first
configuration, called initial state, contained the layout of Areas of potential discomfort in original (top)
the project (the volume envelope prescribed by the City of and modified (bottom) designs

10 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Construction

Pathlines showing wind trajectory colored by velocity magnitude for original (left) and new (right) designs provided increased understanding of wind
behavior near buildings to determine areas of concern.

The ANSYS team determined the amplification factor at


human height for both configurations, illustrating wind
effects such as corner effects near tall buildings and
channeling effects between closely spaced buildings. For
the initial state, a map of the most-restrictive values (in
which AF is greater than critical value) revealed areas where
local wind speed was too high and would lead to
pedestrian discomfort. The modified configuration reduced
areas of discomfort by approximately a factor of three.
The team used ANSYS engineering simulation to
determine airflow between the buildings, and these results
were used to generate velocity vectors to identify specific
points of interest, such as recirculation zones in which
dust could accumulate. The details provided information
that helped in adjusting the building design to avoid these
types of zones. Pathlines of wind trajectory provided
increased understanding of wind behavior near buildings,
including turbulence, and identified areas where potential
problems could occur. Finally, computing pressure con-
tours on the buildings identified areas that would require
special care during the design phase to avoid other con-
cerns, including whistling problems due to insufficient
insulation or damage to the facade.
This project demonstrates the use of engineering
simulation for smart urbanization. Such analysis can
provide comprehensive maps of airflow patterns, local
wind velocities and local turbulence intensities — all
useful in planning urban development projects. Simulation-
driven building design provides information about how
Pressure contours for original and new designs identified areas that
structures affect the environment, which can be an required special care during design to avoid insulation whistling
important part of discussions and decisions from a problems or potential façade damage.
project’s first stages. n

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 11


ENERGY

Image © iStockphoto.com/MichaelJay
Catch the Next Wave
Hydrodynamic simulation helps to deliver two- to three-times
wave power efficiency improvement.
By Bradford S. Lamb, President, and Ken Rhinefrank, Vice President of Research and Development,
Columbia Power Technologies, LLC, Corvalis, U.S.A.

If all the ocean’s energy could be harnessed, it would


produce more than 500 times the global energy consump-
tion. The practical potential for wave energy worldwide
is projected to be between 2 trillion and 4 trillion kilowatt
hours per year. The World Energy Council estimates that
about 10 percent of worldwide energy demand could
realistically be met by harvesting ocean energy.
But wave power is a much less mature technology
than solar or wind power or, especially, fossil fuel.
A tremendous amount of work lies ahead in optimizing the
design of wave power systems. Researchers must
improve efficiency and reduce costs to the point that
these systems can make a major contribution to meeting
global energy requirements.
Columbia Power Technologies (COLUMBIA POWER),
LLC, is attempting to harness this potential by developing
commercially viable and scalable wave power generation
systems. In conjunction with Oregon State University, the
company is working to develop and commercialize
innovative wave energy harvesting devices.
There are several key advantages of wave power:
• Power density: Wave power is much denser than
other renewable energy systems, enabling wave
parks to produce large amounts of power from a
relatively small footprint.
• Predictability: The supply of energy from wave
power can be accurately forecast several days in
advance, enabling utilities to make precise sourcing
plans.
• Constancy: Unlike solar power, which produces
energy only when the sun is shining, ocean swells
are available 24 hours per day.
• Proximity to load centers: Wave energy will not
require substantial buildout of transmission
capacity, since 37 percent of the world’s population
live within 60 miles of a shoreline, and 70 percent
reside within 200 miles. Preparing to test the wave power device

12 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


ENERGY

Wave direction

1 2 3 4 5
COLUMBIA POWER’s wave power system: The wings and vertical spar react to the shape of the passing ocean swell. Each wing is coupled by
a drive shaft to turn its own rotary generator.

The wave power industry, however, faces a major offers both frequency and time domain solutions.
challenge since product developers have much less Frequency domain solutions are faster, which makes them
experience in the design of wave power devices relative ideal for quickly evaluating a large number of shapes,
to other renewable energy systems. Wave power while time domain solutions provide the high level of
companies need to rapidly advance efficiency and reduce accuracy needed to refine to the best shapes in the later
costs of their designs to demonstrate viability to potential stages of the design process.
investors and customers. Other industries have taken COLUMBIA POWER engineers developed an initial
decades or longer to develop technology to the point of concept design in SolidWorks ®, built a prototype and
commercial viability. But the wave power industry does tested it at 1/33 scale in the Tsunami Wave Basin at the
not have that kind of time. To achieve its goals, it needs to Hinsdale Wave Laboratory at Oregon State University. The
rapidly improve designs while conserving limited capital. team used high-resolution cameras to track light-emitting
COLUMBIA POWER is focusing on development of diodes on the buoy, measuring its motion in the waves.
direct-drive systems, which avoid the use of pneumatic Engineers exported the concept design to ANSYS AQWA
and hydraulic conversion steps and their associated software and performed a time domain simulation while
losses. The company believes that direct-drive systems using a wave climate with the same amplitude and
are the future of wave power because they are more frequency as that measured in the wave tank. There was
efficient and reliable as well as easier to maintain. The a very good match between the measurements and
number-one design challenge was to optimize the design predictions from ANSYS AQWA. Since then, engineers
of the buoy to maximize the proportion of wave power have used ANSYS AQWA as their primary design tool to
transferred to the buoy. Relative capture width is a optimize the shape of the fiber-reinforced plastic (FRP) buoy.
dimensionless measure of the
efficiency of the device in
capturing the available energy
of the wave. A relative capture
width of 1 means that the buoy
has captured 100 percent
of available wave energy.
As COLUMBIA POWER
set out to determine the
optimal shape for the buoy,
engineers looked at five
different hydrodynamic simu-
lation software packages. The
company selected ANSYS
AQWA software because of its
ease of use, and tests showed
that it provided a better match
with physical experiments
than did competitive software.
C OLUMB I A PO W ER a l s o
COLUMBIA POWER engineers doubled efficiency of the buoy by using
valued that ANSYS AQWA ANSYS AQWA to optimize its geometry.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 13


ENERGY

COLUMBIA POWER has since evaluated over 350 the power takeoff to the wave climate by changing the
different geometries with ANSYS AQWA in an effort to amount of current produced by the generator, which, in
maximize the relative capture width of the buoy. At the turn, changes the mechanical load placed on the system.
same time, the company worked closely with Ershigs Inc., This makes it possible to consider in a single model the
its structural partner that produces the FRP floats, to effects of different buoy shapes, power takeoff system
explore the manufacturability of various shapes and to designs and control strategies; it also helps to determine
ensure that the final design can be produced at a low the power that would be generated by each approach in a
cost. The company also looked at the survivability and variety of different wave climates.
environmental impact of proposed buoy designs. COLUMBIA POWER recently began using Maxwell
COLUMBIA POWER engineers used a sinusoidal wave electromagnetic simulation software from ANSYS to
shape and a suite of wave frequencies ranging from 2 optimize the design of the generator. Engineers evaluated
seconds to 20 seconds for frequency three different electromagnetic simulation
domain simulations. The response software packages and concluded that
amplitude operators calculated by ANSYS Maxwell was the easiest to use and the
AQWA software were used in a post- most stable. Maxwell is being used to
processing routine written by COLUMBIA analyze the electromagnetic performance of
POWER engineers that calculates the the generator while varying the air gaps
relative torque and speed of the buoy as between the rotor and stator, different
well as the relative capture width. magnet geometries, different magnet
Once they felt that they were close to types, and different types of steel. The over-
an optimal shape for the buoy, COLUMBIA all goal is to maximize the generator’s
POWER engineers moved to time domain energy output while minimizing its cost.
modeling, which makes it possible to As a technology startup with far-from-
evaluate the nonlinear effects of the unlimited funding, COLUMBIA POWER
waves. The team evaluated the shapes must be capital efficient. By focusing its
that had proven best in frequency domain development efforts on simulation and
modeling against a variety of wave using physical testing judiciously as a
climates, including those found at seven verification tool, COLUMBIA POWER is
different coastal locations around the moving forward in the development process
world. At the same time, engineers began much faster than would be possible using
optimizing the power takeoff system that traditional development methods. ANSYS
converts mechanical energy into electrical Maxwell software from ANSYS AQWA and Maxwell simulation software
energy. ANSYS AQWA model results from was used to optimize the enable the company to make its mistakes
generator design.
frequency domain models were post- in the computer, where they are far less
processed in Matlab® Simulink® to incorporate the power expensive than in the ocean. ANSYS AQWA technology,
takeoff reaction torque and to compute power output. in particular, helped to more than double the efficiency of
The ANSYS AQWA time domain models were coupled to COLUMBIA POWER’s wave power system. COLUMBIA
a DLL that simulated both linear and nonlinear power POWER has benefitted from the excellent technical
takeoff operation. The DLL for the power takeoff model support and productive training sessions provided by
was developed in Matlab Real Time®. Engineers used the ANSYS. As a result, the company is on track to soon
output from ANSYS AQWA to drive a numerical model deploy the first ocean demonstration of its technology in
developed in Simulink that simulates the power takeoff Puget Sound. n
system and control strategy. The control strategy tunes

14 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Energy

Reacting to Emissions
Fluids simulation helps to speed up research into chemical-looping
combustion capable of reducing fossil fuel emissions.
By Schalk Cloete and Shahriar Amini, Flow Technology Research Group, SINTEF Materials and Chemistry, Trondheim, Norway

The need to reduce CO2 emissions to restrain climate N2 Metal oxide CO2 +
H2O
change has never been more urgent. Anthropogenic CO2 is
mainly generated in the combustion of fossil fuels, and
these fuels are expected to provide a very large percentage Oxidizer Reducer
of overall world energy consumption for the next several QO

decades. One thing is clear: Emissions must be lowered.


An accepted solution is to separate and sequester the Qhot ηrev Qcold
CO 2 emitted by fuel combustion. Chemical-looping
combustion (CLC) is one of the most promising tech- W

nologies to carry out CO2 capture at a low cost.


CLC is a combustion process that avoids direct
mixing of fuel and combustion air. The method uses two
fluidized bed reactors and circulating metal oxide that is Air Fuel

oxidized in an air reactor and reduced in a fuel reactor Metal


to provide the oxygen required for the fuel. Pure CO 2
Chemical-looping combustion process
is obtained in the fuel reactor exhaust stream after
condensation of water without the need for further gas made much easier by full access to all-important models
separation. CLC will achieve significant CO 2 capture by employing user-defined functions (UDF) compiled within
at a reduced cost when compared to other current tech- the stable, parallelized ANSYS FLUENT solver.
nologies, including post-combustion amine scrubbing. Heterogeneous reactions, alternate drag laws and wall
The delay in timely commercialization of CLC functions for dilute granular flows were implemented
technology is primarily due to lack of understanding of through UDF along with several other variables, such as
reactive multiphase flows in the fluidized beds used in adjusting pressure drop over a periodic section to maintain
CLC systems. Design and scaleup of CLC reactors are a constant superficial gas velocity. Studies were conducted
very complex. Advanced modeling techniques on coarse graining, implementing filtered
are required to capture the intricate coupling drag, solids viscosity and solids pressure
between the complex reactor hydrodynamics, formulations by means of UDF.
heterogeneous reaction kinetics and heat One requirement for modeling the
transfer. reactive gas–solids flows in CLC processes
With the advent of increased computational is that the software predicts formation of
capabilities, the fundamental modeling frame- the particle substructures (clusters) that
work of computational fluid dynamics (CFD) is often occur in these systems. These
emerging as a very promising additional tool structures appear as gas bubbles in dense
for modeling reactive multiphase transport. regions and particle clusters in dilute regions
Once the model has been validated, CFD can of the fluidized bed reactor, and they
be used for accurate design and scaleup of the influence everything happening in the reactor.
process. A team at the Flow Technology group Fine meshes and small time steps are
at SINTEF, the largest independent research required to resolve particle structures, but a
organization in Scandinavia, has been using price in computational time has to be paid
ANSYS FLUENT software for this purpose in a to capture the physics of the system.
project led by Shahriar Amini. The well- These particle structures increase the slip
established granular flow modeling framework Simulation shows both between particles and fluidizing gas. This
available within the software has been central the bubble (bottom half) effect is analogous to the clustering of tiny
and cluster (top half)
to the progress made in this project. structures occurring in
mist droplets into larger raindrops. A mist of
Fundamental model development also was fluidized beds. microscopic droplets does not fall and is very

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 15


Energy

easily blown away by the wind, reactor domain. Since the reac-
whereas a big raindrop certainly tion occurring on the cluster sur-
falls with considerable force. face contributes only a small
The same happens in a fluidized amount to the global reaction
bed reactor: The tiny particles rate, the correct description of
conglomerate into larger this interface becomes less
clusters and, therefore, fall much important.
more readily. If this clustering is Proper resolution of particle
not modeled correctly, the simu- structures along with achieving
lation will predict a bed hydrodynamic and reaction
behaving like a mist, but it kinetic grid independence require
should actually behave like a engineers to pay particular
thundershower. Clearly, this will attention to grid size compared
A particle cluster (left) influences the concentration of
lead to substantial errors. the reacting gas (right).
to particle diameter. To achieve
In addition, it is believed that hydrodynamic grid independence
particle structures influence reactions. High reaction rates for denser beds, the grid size can be 25 particle diameters
are observed in regions with high volume fractions of or even higher, while the fine structures formed in risers
particles and reacting gas concentrations. A dense need grid sizes of a maximum of 15 particle diameters,
particle cluster represents a region in which lots of surface provided high-order spatial and temporal discretization is
area is available for reaction, and this significantly used. Reaction kinetic grid independence works the
increases the reaction rate. However, the consequence is opposite way. It is very sensitive in dense beds since the
that the reacting gas inside the cluster is used up much volume fraction gradient on a bubble is so large. For very
faster, thereby slowing down the reaction. Clustering is high reaction rates, the grid size can be 10 particle diam-
actually an unwanted phenomenon in fluidized bed eters or lower, whereas for 50-times lower reaction rate
reactors because it slows down the reaction rate overall the grid size can be 25 particle diameters. Reaction kinetic
by concentrating all the solids in areas with a low amount grid independence in risers is slightly more forgiving, since
of reacting gas. Incorrect cluster modeling, therefore, the clusters can be more dilute and normally coincide with
will predict better, but unfortunately very wrong, hydrodynamic grid independence.
reactor performance. Interestingly, simulation results show that the use of
The team at SINTEF has accepted the challenge to a highly reactive particle will not improve reactor
contribute fundamental knowledge in the field of simu- performance as much as might be expected. In the highly
lation of structure resolution on reactor performance. reactive case, no reacting gas species is allowed to
Results have shown that the degree to which the clusters penetrate into the cluster. A large percentage of the
have to be resolved (and therefore the particles available for reaction, there-
computational cost of the simulation) fore, is being wasted inside the dense
depends largely on the reactivity of the clusters where no reacting gas species
particle used in the reactor. When a is present. When the particle with a low
highly reactive particle is employed, the reactivity is used, fuel gas is available
reaction between the fuel gas and a throughout the reactor, and all particles
dense cluster is almost instantaneous are involved in the reaction. To quantify
and occurs almost exclusively on the this concept, simulations have been
surface of the cluster. In this case, used to show that a 50-times decrease
the most important phenomena to be in particle reactivity decreases the
correctly modeled are the area of the overall reaction rate by a factor of
cluster on which the reaction can occur two to three, depending on the
and the transport of the reacting gas fluidization velocity.
species to this interface. Any slight error Information gained from this study
in the position or sharpness of the is currently being used at SINTEF to
cluster surface will lead to significant perform fluid simulation as a design
modeling errors. and optimization tool for CLC systems.
When a less reactive particle is used Fundamental insights offered by these
in CFD simulations, however, the The interaction between particle clusters models make substantial contributions
resolution of the cluster interface and reacting gas using a particle with toward identifying and optimizing
high reactivity (two images on the left)
becomes less important. In this case, and low reactivity (two images on the important design parameters to
the reaction rate is slow enough to allow right). In each pair of images, the one accelerate the development of this
on the left shows the particle volume
reacting gas to penetrate into the parti- fraction, and the one on the right shows important CO2 capture technology. n
cle clusters and react throughout the the reacting gas concentration.

16 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


construction

Meeting Green Building


Design Goals with
Engineering Simulation
Simulation is driving innovation in HVAC design for an assembly hall
in Saudi Arabia.
By Sharbel Haber, Senior Mechanical Engineer, Balsam Nehme, Mechanical Engineering
and Adnan Akhdar, Mechanical Engineer, Dar Al-Handasah, Beirut, Lebanon

With the ever-increasing demands for sustainable In the


buildings, engineers are developing more complex and assembly
diversified designs to reduce loads, boost efficiency and hall of the
utilize renewable resources. Fluid dynamics simulations convention center,
have proven to be a powerful and effective tool, providing an underfloor air
flexible solutions in increasingly complex and demanding distribution (UFAD)
projects. At engineering firm Dar Al-Handasah, these system was coupled
simulations are extensively used as an optimization and with a conventional ceiling Full-scale 3-D model of the
validation tool at an early phase in the design process, air supply. The assembly convention center assembly hall
since simulation supports implementation of innovative hall that includes the stage
designs and energy-saving measures geared toward and seating area consists of four levels (basement, ground,
decreasing the overall facility’s energy costs while main- mezzanine and first floor) interconnected via one air
taining or improving occupant comfort. continuum, with all levels occupied. Dar Al-Handasah
Dar Al-Handasah (Shair and Partners) has been a was charged with reducing the HVAC system energy
pioneering force in the planning, design and implementation consumption while ensuring audience comfort: The total
of development projects in the Middle East, Africa and Asia airflow supplied by the hybrid cooling system should cool
since it was founded in 1956. Today, Dar Al-Handasah is the occupied zones but only temper the upper regions
one of the largest engineering and design firms in the world. where maintenance catwalks are located. The distribution
A typical example of building for energy efficiency is the of air supply outlets needed to be optimized to ensure
company’s recent design for the 31,000-square-meter con- proper air delivery in the 24-meter-high assembly hall while
vention center at Princess Noura Bint AbdulRahman avoiding disturbance of the thermal stratification of
University for Women in Riyadh, Saudi Arabia. the air, a key energy-saving measure. Fluid dynamics

Rendering of the convention center at Princess Noura Bint AbdulRahman University for Women in Riyadh. The center assembly hall was studied
to reduce energy costs while maintaining comfort.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 17


construction

simulations with ANSYS design. Spot cooling was


FLUENT software were used to balance the high
fundamental in optimizing heat emitted by the cat-
the UFAD system to maxi- walk’s equipment so that
mize energy savings and reduce disturbances would not be
energy costs. introduced into the upper
To analyze the performance of hot stratified air layer.
the proposed system, engineers at To capture the rising move-
Dar Al-Handasah built a full-scale, ment of buoyant hot air from
three-dimensional virtual model of the occupied zones, the
the assembly hall, representing the team added air returns
complex geometry of the above the catwalks with
building. They accurately an exhaust fan installed
modeled 902 air diffusers at the highest elevation of
beneath the audience seats in the stage. This fan also
the hall and 62 ceiling flow bars assisted in exhausting any
distributed throughout the different contaminants.
levels. For greater accuracy, the Fluid flow simulations
team then meshed the 3-D model have become an instru-
with local mesh refinements applied mental tool in supporting
in the occupied zones and near the Original design scenario (top) and optimized design model t h e c o m p a n y ’s d e s i g n
(bottom) for temperature distribution across the convention
air outlets. Engineers conducted center assembly hall process through accurate
steady-state simulations using the prediction of thermal
ANSYS FLUENT solver to optimize comfort conditions, design
the performance of the proposed hybrid air conditioning validation enabling design optimization, and energy
system to produce an environment that complies with reduction for the HVAC building systems. Furthermore,
the required comfort conditions. In particular, velocity and the proven breadth and depth of advanced fluid dynamics
temperature distributions were generated and air distri- modeling capabilities from ANSYS have allowed the
bution refined so that no disturbances occurred in the company to tackle a wide range of complex problems
hot stratified region. ranging from HVAC and smoke simulations to dispersion
Initially, the flow simulation indicated a potential for modeling and pumping stations simulations. Dar
improving the temperature distribution at various levels. In Al-Handasah, with the support of ANSYS channel partner
fact, high temperatures were observed at the first-floor- Fluid Codes Ltd for the Middle East, continues to explore
level seating area (around 31 degrees C) and at the stage opportunities to incorporate state-of-the-art tools in
area (around 25 degrees C), implying insufficient supply air building design to continuously improve design quality
flow to these zones. The team observed very high temp- while exceeding client expectations. Fluid dynamics
erature (around 40 degrees C) in the core volume mainly software from ANSYS helps engineers to optimize HVAC
due to high heat loads from the equipment on the catwalk. designs and meet the ongoing challenge of developing
In addition, the mezzanine-level seating area was over- models that are energy efficient, sustainable and compliant
cooled, with average temperature around 21 degrees C, with standards. n
indicating that the supply air flow delivered to that zone
could be reduced.
Using the results from the initial fluid dynamics
simulation, the engineers were able to visualize the airflow
behavior inside the high-ceilinged hall and devise improved
airflow delivery parameters, which were validated in a
second simulation. Specifically, the supply air flow in the
first-floor seating area was increased and directed
at predefined angles, leading to better air delivery and
accordingly lower temperature. Additionally, the team found
the side flow bar diffusers at the balconies were inducing
Optimized design model showing flow pattern and velocity distribution
significant disturbances to the hot stratified air layer across the assembly hall. Simulation assisted the designers in meeting
and, thus, engineers removed those diffusers from the energy efficiency goals.

18 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


offshore

Sensor float fluid–structure interaction (FSI) transient response Sensor float bow nosing down due to flow on sensor below float,
to current flow deck awash

Don’t Rock the Float!


Fluid–structure interaction allows designers to assess
impact of waves on freshwater and offshore systems.
By Richard Grant, President, Grantec Engineering Consultants, Inc., Halifax, Canada

On a recent project commissioned by Environment most powerful and trusted fluids and structures simulation
Canada, Grantec Engineering Consultants, Inc. was tools. With its multi-field solver, the ANSYS FSI solution
tasked with developing a water quality monitoring float provided Grantec’s team with a bidirectional capability for
designed to carry a sensor for capturing environmental time-transient or steady-state analysis with moving or
data. The float plays a role similar to — and looks deforming geometry. Using ANSYS Multiphysics software,
somewhat like — a catamaran, though it is designed to be the Grantec engineers were thus able to evaluate both the
moored rather than driven by an engine or sails. The goal structural part of the analysis and the fluid flow solution
of the analysis was to minimize drag and ensure stability with just a single tool.
of the float as well as to develop specifications for the In the original float design, the team modeled the float
mooring system and structure. To meet this goal, Grantec and sensor as a flow obstruction, which accounted for the
used multiphysics simulation software from ANSYS to flow currents and wave loading on the float as well as
determine the fluid–structure interaction (FSI) by modeling buoyancy forces. They then evaluated the development of
the float and sensor under a wide range of water current bow and stern waves that result from the resistance of the
and wave conditions. hull to fluid flow, just as with the hull of a ship. The
Based in the maritime province of Nova Scotia on the software duplicated the vertical heaving and angular
east coast of Canada, Grantec and its engineers have an pitching of the float in response to different wave and
extensive background in both structural and fluids analysis current conditions. The impact forces from the waves
helping customers in the defense, offshore, marine, calculated in the fluid simulation were automatically
manufacturing, energy and aquaculture fields advance passed back to the structural model to more accurately
new designs and systems. More recently, however, simulate the stresses and deformations on the hull.
Grantec has often faced the challenge of how to combine Though they have little effect on fluid flow, the stresses
these two analyses that have historically been performed are important because they make it possible to optimize
separately. Previously, when the interaction between fluid the design of the hull to a much higher level than would
and structure was critical, Grantec’s engineers needed to be possible without them.
enter the results from the fluid dynamics software manually With the FSI solution from ANSYS, Grantec evaluated
into the structural analysis software and vice versa. In the performance of a wide range of hull profiles and mass
contrast, ANSYS offers a solution integrating several of its distributions under different flow conditions, and it took

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 19


offshore

advantage of parallel processing to accommodate larger


models more efficiently than using a single-machine
environment. In the initial series of designs studied, the
sensor was fixed to the stern of the float and extended
vertically into the water. The FSI results for these designs
showed the force exerted by water currents on the sensor
combined with the bow wave tended to push the bow of
the float underwater in faster currents. It was not practical
to solve this problem by simply changing the hull design,
so the team tried a hinged connection between the sensor
and the float to reduce the load transmitted from the
sensor to the float. The hinged sensor, however, greatly
increased the complexity of the simulation analysis.
Grantec addressed the new challenge of the hinged-
sensor design by modeling the float with the sensor fixed
in different hinge positions using the immersed pipe
element in the structural portion of ANSYS Multiphysics
software. Unlike the extensive approach used for the non-
hinge designs, this new method provided a more simplified
way to perform FSI analysis. With the immersed pipe
element, the team applied wave and current loading to
the structural model without the computational load
involved in coupling it to a full fluid dynamics analysis. In
the future, Grantec plans to use a moving mesh to perform
a more complete FSI analysis including full fluid dynamics Waves washing over top of gravity-based structure of offshore platform
simulation that will evaluate the motion of the hinge in (waves traveling to the right)

response to hydrodynamic forces.


Beyond its studies of water quality monitoring floats,
the company has done extensive work with engineering
simulation to help create safer and more structurally
sound offshore structures and systems. Grantec’s
engineers have also used the ANSYS Multphysics solution
to assess gravity-based structures (GBSs) used to protect
offshore oil drilling and production platforms from
icebergs. GBSs rely on weight to secure them to the
seabed, which eliminates the need for pilings in hard
seabeds. Concrete GBSs are typically built with huge
ballast tanks so they can be floated to the site and, once
in position, sunk by filling the tanks with water. The
Grantec team used FSI from ANSYS to simulate wave
loading a GBS including the effects of massive waves
from storms — also known as green water — coming over
its top.
The company believes that its investment in ANSYS
Multiphysics software has made a significant addition
to its analytical capabilities. Clients seek out Grantec
because of its track record in performing advanced
engineering to solve very complex problems. ANSYS
technology has helped put another tool in the Grantec
toolbox that makes it easier to address design challenges
that just a few years ago would have been much more
Finite element mesh (top) and contours of stress (bottom) on a half difficult. n
model of the sensor float. The FSI analysis was performed to look at
the effect of a fixed flexible boom on the float.

20 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Invented the Supercomputer,
Reinvented the Workstation
The Cray CX1 brings high-return, low-risk
benefits to design productivity and
efficiency for ANSYS® FLUENT®, ANSYS®
Mechanical™ , HFSS™ , and more.

Visit the Cray and ANSYS


Partner Media Center at
www.cray.com/reinvent

www.cray.com/reinvent

The leader in
Cluster, Grid and Cloud
Management Software

More Power.
More Possibilities.
Integrate. Simulate. Dominate.

When ANSYS applications are powered by Platform HPC clusters, your engineering
team can deliver more designs, more accurately, and faster. Run more complex and
accurate simulations than previously possible to improve product quality and reduce
time to market. Empower yourself.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 21


Energy: Wind

Energizing the Wind Industry


Increased complexities require a system-level approach
in designing and evaluating wind turbines.
By Ahmad Haidari, Global Industry Director for Process, Energy and Power, ANSYS, Inc.

Wind energy projects around the globe — from small installations to very large wind
farms — have a common goal: to reduce unit energy cost while improving reliability.
From a business perspective, technology contributes to viability by influencing
efficient wind turbine design, manufacture, deployment and operation. Whether
the application is an onshore, offshore or far-shore installation, advancements in
science and engineering will contribute to the industry’s success, especially through
capabilities related to aerodynamic design, material science, structural design, Speed sensor design

electronic mechanical control, site selection and farm layout.


Wind turbines and wind energy projects are becoming increasingly more
complex, so they must operate dependably at levels unimaginable a few years ago.
Installations of very large wind turbines in offshore and floating configurations are a
major technological achievement. Energy companies hope to design, install, and
efficiently and reliably operate superstructures whose wind blade spans are over
50 meters and subject to wave and wind loading at different angles of attack.
Historically, wind energy companies have used engineering simulation software
as a point solution, used only to simulate a specific design aspect or analyze a
Transformer design
component. Successful application of ANSYS solutions ranges across the wind
energy industry, including:
• Aerodynamic design: thrust coefficient, blade structural integrity, ultimate
loads and fatigue, noise predication, wind gust fluid–structure interaction, bird
strike, icing, boundary layer transition, near-wake and far-field studies
• Structural design: tower and rotor structural integrity/safety, power conversion
efficiency, installation cost and maintenance, offshore transport and installation
• Component design: blades, gearboxes and bearings, generators, nacelles,
rotors, drivers, motors, electronics cooling
• Site selection and farm layout: maximum project potential, power output Electric machine
(both peak and average), wind loads, fatigue analysis
• Turbine placement: variable terrain, roughness, forestry, multiple wake effects,
buildings and setbacks
• Electromechanical system: electrical machines, variable-speed control
systems, transformers, power electronics, power distribution, sensor and
actuator design
• Blade manufacturing

Today’s increased complexities require a system-level approach in designing


wind turbines and evaluating performance based on real-world conditions.
Advances in engineering simulation software increasingly make this possible: For Power electronics design
example, the ANSYS Workbench environment is designed with capabilities that
enable modeling entire wind turbine systems. Its value is further enhanced through
advanced solver functionality including turbulence transition models, advanced
contact models, multiphysics capabilities, composites tools, high-performance
computing and the flexibility to connect to third-party software for wind turbine
blade manufacturing or aero-elasticity calculations.
Engineers can perform electromechanical system-level analysis using Simplorer
software, electromagnetic analysis on electric machines and drives with Maxwell,
wind power analysis via ANSYS CFD, and stress and modal analysis using ANSYS Generator and
shaft design
Mechanical. By leveraging high-level integration and advanced capabilities,

22 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Energy: Wind

Wind Farm and Power Distribution

Power distribution analysis Site selection, land and sea Wind farm configuration for Design of floating
optimal power generation offshore wind turbine

engineers over time are extending their once-simplified With increased demand for wind energy, engineers will
simulations to include additional details in overall wind face additional complexities, such as even-larger turbine
turbine design — enabling small efficiency gains, blades that will be installed farther offshore and in harsher
important in an industry in which a minute efficiency/ environments. Wind farm site selection must continue to
performance gain can translate into much larger electricity reduce risk and overcome proximity and environmental
production, reduced downtime and greater project profit- concerns. New powertrains, lighter towers, multi-access
ability. Such details can also improve reliability and enable turbines, floating platforms and quieter machines will be
better wind energy project operation. developed. The industry will innovate to meet the

Some images courtesy CENER, IMPSA, REpower Systems AG, Trane Inc.,
Already, there are many exciting examples of the challenges of increased safety and reliability, improved
expanding use of engineering simulation throughout remote monitoring, reduced system maintenance and
the wind energy supply chain. Three separate wind energy regulatory concerns. ANSYS is keeping pace by providing

Xanthus Energy Ltd. and iStockphoto.com/Tore Johannesen.


applications are presented in this issue of ANSYS high-fidelity integrated, advanced capabilities that meet
Advantage — and each highlights the breadth of single-physics needs as well as system-level and multi-
ANSYS solutions. Without simulation capabilities, the disciplinary requirements of the wind energy industry. n
projects presented may not have been as successful.
Software from ANSYS meets the challenges of individual applications as well as of complete systems in the wind energy industry.

Blade design

Rotor sizing
and acoustics
Tower design and fluid
structure interaction

Offshore installation
and certification

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 23


Energy: Wind

More Power
to You
Simulation helps Indar to design
one of the world’s highest-
efficiency permanent magnet
wind turbine generators.
By Jon Vaquerizo, Project Manager, and Xabier Calvo, Technical Indar high-speed permanent magnet generator with
Manager, Indar Electric, S.L., Beasain Guipuzkoa, Spain air–water cooling option

Wind power is the world’s fastest-growing energy source, Indar Electric was founded in 1940 as a manufacturer
with 37.5 gigawatts of installed capacity added in 2009. of small electric motors. In 1997, it became part of
The Global Wind Energy Council expects this resource to Ingeteam, a Spanish renewable energy company that
grow by 160 percent over the five-year period ending currently holds about a 15 percent global market share for
2014. One rising trend is permanent magnet generators wind power components. Indar produces a wide range of
(PMGs), as they offer higher efficiency and design flexi- generator concepts, including more-traditional double-fed
bility. Indar Electric, S.L., set out to develop a 2.5 MW PMG induction generators (DFIG) and newer PMGs. PMGs gen-
for wind power applications with the ambitious target of erally offer higher efficiency at rated load and even more
achieving an unprecedented 97.7 percent level of effi- at partial loads, since the permanent magnets eliminate
ciency at rated load in converting mechanical to electrical the need for rotor windings that, in turn, remove rotor
energy using a permanent magnet generator. Another goal ohmic losses. PMGs also eliminate the need for brushes,
was to increase performance and efficiency at partial which reduces possible problems and maintenance
loads, because wind turbines often run at partial load. needs.
Traditional build-and-test methods could not achieve The Indar design team faced several major challenges
these goals in a reasonable amount of time. Thus, Indar in developing its newest PMG. Achieving high efficiency
applied electromagnetic field and fluid flow simulation to was the overarching goal, but there were a number of other
facilitate the process. targets that had to be simultaneously achieved for reliable
operation. Cogging torque, caused by the interaction
between the rotor’s permanent mag-
1250
1000
nets and the slots on the stator core,
750 had to be reduced to 0.1 percent of
500 overall torque. Voltage harmonics
250
(THD) in the output had to be kept
Volts

0
0.005 0.01 0.015 0.02 below 0.5 percent. For the cooling
-250
-500 system, a demanding goal was set
-750 to maintain magnet temperature
-1000 below 100 C to assure good per-
-1250
Time formance of the magnets over a
Actual value of 1st order is 99.9 20-year lifetime.
0.7
Indar engineers used Maxwell
Magnetic flux density level in the
0.6 low-frequency electromagnetic field
stator (left) identifies areas of high 0.5 simulation software from ANSYS to
losses. The generated voltage (upper
Voltage (%)

right) and fast Fourier transform (FFT)


0.4 evaluate the effect of different
(lower right) results for a no-load 0.3 geometries and magnet properties
condition are also shown.
0.2 on the electromagnetic performance
0.1 of the generator. Well-known basic
0.0 equations were used to develop
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Order of Harmonics a preliminary generator design.

24 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Energy: Wind

250

200
100
150

50 100

50
Torque (%)

Current (%)
0
0

-50
-50
-100

-100 -150

-200
-150
0.95 1 1.05 1.1 1.15 1.2 -250
0.95 1 1.05 1.1 1.15 1.2
Time (s) Time (s)

Magnet behavior in a three-phase short circuit as predicted by Maxwell

Engineers first created a 2-D and, later, a 3-D model of the In the full-load simulation, engineers looked at the
generator, relying primarily on manufacturing drawings to input required by Ingeteam frequency converters for the
reproduce the geometry and material properties of rotor available switching frequencies to achieve nominal torque,
and stator laminations and coils. The time step of the high current and low losses. The team examined induction
simulation was adjusted to match the rotating speed of levels in the stator because of their important effect on
the generator and the number of poles in the permanent efficiency. Though high induction levels make it possible
magnet. Engineers simulated the performance of to reduce the size of the generator, they also increase iron
the proposed design under no-load, full-load and short- losses. Maxwell results showed the distribution of losses
circuit conditions. over the geometry of the stator, providing guidance for
The results of the simulation included the voltage design changes to improve efficiency. Indar engineers
wave form produced by the generator, a measurement continually modified the design, attempting to reduce
that was compared to the design requirement so that losses in the stator copper, mechanical losses, and losses
harmonic levels could be evaluated. The voltage fast created due to switching frequency when working with
Fourier transform output capability in Maxwell provides frequency converters — all while achieving the other
the voltage at different frequencies, making it design requirements.
straightforward to calculate the harmonic levels for a
particular design.
The behavior of the design in the event of a short
circuit was another important consideration. Short circuits
may be caused by mechanical failure in the generator,
insulation breakdown or power converter malfunction.
Engineers studied the magnetic field generated in each
area of the permanent magnet in a short circuit, expecting
to ensure that the magnet had the right properties to avoid
any damage. Generally when designing a PMG, Indar
engineers consider a wide range of factors, including the
contribution of magnet temperature, rotational speed,
switching frequency and short-circuit (two-phase and Flow speed and temperatures through the tubes in the air–air
three-phase) performance to achieve ideal magnet cooling system

behavior for the entire lifetime of the generator circuit. Because of the magnet field’s high strength even when
the PMG is not rotating, Indar engineers simulated the
process of assembling and balancing the rotor, ensuring it
could be safely accomplished. They determined the level of
magnetic forces generated while inserting the rotor, which
made it possible to specify assembly tools that could
withstand these forces. The generator rotor is balanced by
placing it on pedestals instrumented with accelerators
that detect forces generated by imbalance in the rotor.
The electromagnetic field generated by the rotor during
this process was simulated with Maxwell to ensure
Electromagnetic simulation of the balancing operation and inserting the
it did not interfere with the cables carrying the
rotor in the stator with a crane accelerometer signals.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 25


Energy: Wind

10
prototype was built. Pervasive parameter management was
9 performed quickly and easily by changing design para-
8 meters. These parameters were propagated throughout the
7 entire design system, from CAD model through meshing
6 and boundary conditions to generation of updated results.
Air Flow (%)

5
The next step was building and testing a real-scale
prototype to verify the simulation results and to ensure the
4
Option 3 generator’s functionality and life expectancy. Two types of
3 Option 2
Option 1 testing were performed at the test bench: generator full-
2
load homologation testing to certify generator performance
1 and durability testing to verify its reliability over time. The
0 electrical and thermal measurements of the physical proto-
Generator Length type matched up very well with the simulation results. For
Fluid dynamics results show air flow variations over the length of the example, the maximum deviation from the voltage shape
generator using three different cooling system options. The options
involve modifying the geometry of the slot, windings and magnets. prediction to the measured values was 0.1 percent.
The measured efficiency of the new generator was 97.86
Indar engineers simultaneously studied the gen- percent, higher than the design target of 97.7 percent, and
erator’s cooling system because of the interaction between nearly exactly what was predicted by the simulation. The
electrical and thermal performance. The temperature of the rating is one of the highest levels of efficiency for any per-
magnet plays an important role in its ability to resist manent magnet generator on the market. Simulation made
demagnetization, so improvements in cooling performance it possible to achieve this challenging performance goal in
can increase the magnet’s ability to handle a short circuit. less than half the time that would have been required using
Optimization of the cooling circuit helps to improve conventional build-and-test methods. The simulation pre-
efficiency by reducing mechanical and cooling losses. dictions correlated well with physical testing, providing
To optimize the cooling circuit, engineers used ANSYS confidence that Indar can use simulation to optimize its
FLUENT fluid dynamics software to perform a detailed products to deliver high performance under the most
study of fluid flow and heat transfer in and around the demanding conditions. n
generator. Meshing was a challenge because of the
difference in scale between the small 5 mm to 10 mm air 8.0
Max. Deviation = 1.32%
gap between the rotor and stator, where accuracy was 7.5 Mid. Deviation = 0.5%
critical, and the large overall 1 meter length of the generator 7.0
Air Flow (%)

and cooling system. To minimize computational time, 3-D 6.5


6.0
steady-state analysis was used during the majority of the
5.5
design process, and the model size was reduced by using
5.0 Simulated
axial symmetry and periodic conditions. Fluid dynamics 4.5 Measured
results included the local heat transfer coefficients, air flow 4.0

velocity at every point in the machine circuit, pressure Generator Axial Length

drop of the air circuit through the generator, generator


temperature and thermal profile, and magnet working
temperature. Engineers used these results as a guide in
Temperature

decreasing temperature hotspots by reducing variations


in cooling over the length of the generator.
Simulated
Using software from ANSYS helped Indar to easily Measured
ex plore multiple automated parametric design
variations. The electromagnetic and fluid flow simulations
Generator Axial Length
provided far more diagnostic information than was avail-
able from physical testing. Simulation provides results for The difference between the simulated and measured air flow rates
across the length of the generator was a maximum of 1.32 percent. The
any output at any point in the computational domain, while maximum difference in calculated versus measured temperatures was
physical testing provides results only at locations where it only +/– 3 degrees C for the stator and +/– 3 degrees C for the rotor
magnets. Software from ANSYS provided very good accuracy, which is
is practical to locate sensors. Engineers were able to iterate essential for these applications and for Indar engineers to trust the new
to a design that met all their specifications long before a design variations.

26 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Energy: wind

Where the Wind Blows


Engineering simulation plays a role in getting the most power from
wind farms by predicting the best available locations.
By Ian Jones, ANSYS Fellow, and Christiane Montavon, Senior Technical Services Professional, ANSYS, Inc.,
and Daniel Cabezón, Wind Specialist, National Renewable Energy Centre (CENER), Sarriguren, Spain

Driven by the need to change the power generation energy include visual intrusion, noise, the effect on nature, and
mix, the use of wind turbines is increasing significantly disturbances caused by construction and operation.
around the world. The Global Wind Energy Council Furthermore, wind farms affect radar signals, a
(GWEC) predicts the global market for wind turbines will condition that is forcing turbines away from exposed
grow from 94 GW in 2007 to 296 GW of total installed hilltops to locations that are out of sight of antenna
capacity by 2012. Depending on the increase in electricity installations. Consequently, wind farm developers are
demand, wind power could supply 11.5 percent to considering sites not necessarily ideal for wind power
12.3 percent of global electricity demand in 2020, generation: locations away from hilltops, in regions where
according to GWEC, and between 18.8 percent and the flow can recirculate and turbulence can be high, and
21.8 percent in 2030. in heavily forested areas. Packing turbines into such
However, the uptake for land-based installations is not constrained sites means that wake effects can have a
as great as it could be. Some of this is explained by significant impact on the performance of downstream
planning considerations: People do not want to have turbines, reducing power production and considerably
wind turbines sited near their homes. Typical objections increasing fatigue loading.

Black Law site


Courtesy Siemens press.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 27


ENergy: Wind

To overcome the shortage of land sites, interest is


Black Law Wind Farm growing in offshore turbine sites, where low ambient
A typical wind farm example is Black Law, in central turbulence is expected to result in reduced rates of wake
Scotland, operated by ScottishPower Renewables. recovery. To minimize wake losses and optimize energy
The site contains 54 2.3 MW Siemens turbines and production, it is vital to improve confidence in the ability
occupies a former open-cast coal site. In January of models to accurately capture these effects.
2006, it was the largest operating wind farm in the No matter the location, there is considerable
United Kingdom. reluctance to invest in expensive wind farms without a
With relatively small height variations across the guaranteed return on investment (ROI). Using current
wind farm (approximately 170 meters), the complex- prediction tools, developers are unable to provide
ity on this site is less related to topography and sufficiently reliable estimates of projected wind power
more associated with significant forestry and wake from challenging sites, failing to satisfy investment criteria.
effects. ScottishPower Renewables and ANSYS Developers are now looking for ways to make wind power
have carried out a collaborative study on this wind estimation more reliable for non-ideal sites.
farm to understand the interactions between the Turbine blades must withstand significant stresses
turbines, the terrain and the forestry, to assess how caused by strong rotation and wind buffeting. There is
well fluid dynamics would perform for this site, and considerable interest in extending turbine life and
to use the lessons learned in developing new sites. reducing operational costs. Understanding the wind
A comparison of ANSYS results and measurements regime and turbulence levels enables greater reliability in
of the annual average power for the Black Law site assessing suitable turbine types. Under the International
shows that the software predicts performance well. Electrotechnical Commission (IEC) (the recognized
The results also indicate which turbines are under- international body for standards development activities)
performing, for example, because of their location turbine classification system, the choice of turbine type is
in the forest. determined by the levels of turbulence intensity likely to
be experienced at the site. For example a Class IIIB wind
turbine, designed for lower wind speed (7.5 m/s at hub
height) and lower turbulence intensity (16 percent), will be
subject to lower loads than a Class IA or IIA wind turbine.

“This type of simulation is useful


in highlighting features not picked

Map of Black Law Ordinance Survey® Crown Copyright 2008,


up in other engineering models
license number 100048580
and in bringing to light
the circumstances under which
standard models fail to capture
Normalized Power

the detail required. Therefore, it


has a key role to play in managing
the risks associated with modeling
Turbine Number complex wind farms.”
(Highest power on left)

Comparison of average wind power on Black Law, taken – Callum Strachan, ScottishPower Renewables
from operational data and ANSYS simulations, shows close
agreement — giving ScottishPower Renewables increased
confidence in using fluid dynamics simulation for future site
application.

28 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Energy: wind

A Class IIIB machine can therefore be larger both in terms


of rotor diameter and hub height to capture a larger Alaiz Wind Farm Test Site
portion of the available energy which will translate in The Alaiz hill, located in northern Spain, is situated
higher capacity factors compared to Class I and II wind in an east–west direction. The site is 4 kilometers
turbines. However, they cannot be used at sites that long and approximately 1,050 meters high, and it
experience a high level of turbulence intensity. comprises very rugged terrain. This hill has been
Understanding turbulence allows a greater choice of used as a test site for gathering wind farm data and
turbine types and enables matching the type to the evaluating different simulation methods. Detailed
site conditions. calculations have been carried out at CENER
The use of ANSYS CFD solutions has increased employing ANSYS CFD software as well as using
significantly for modeling complex flows that arise in wind an industry-standard linearized model. Comparisons
farms as a result of the terrain, forest and wake effects showed significantly lower errors in the predicted
between turbines. Now a customized ANSYS tool helps wind speed for the fluid flow simulation, on
designers to understand wind farm efficiency for complex average about 1.75 percent, compared with about
flows and overcome technology limitations: WindModeller. 5.4 percent for the linear models. This demonstrates
The solution assists developers who don’t have the real benefits that fluid dynamics simulation
substantial CFD expertise to benefit from automated can provide.
analysis of complex flows. Advanced users can further
customize the setup and write post-processing scripts to
capture their own expertise.
An add-on tool for ANSYS CFD — CFDWind — was
created by El Centro Nacional de Energías Renovable
(CENER, or National Renewable Energy Centre), a tech-
nology center in Spain that specializes in applied research,
development and promotion of renewable energies. The
tool contains advanced models for atmospheric boundary
layers, enabling developers to take into account complex
climatic conditions when developing projects. Fluid
dynamics tools from ANSYS provide greater understanding
of the wide variety of factors that need to be considered
when designing wind farms. The flexible nature of the
software provides an excellent platform for customization
specific to this industry. n
Simulation results showing velocity vectors colored by
References velocity magnitude for Alaiz wind farm test site
1. http://www.wind-energy-the-facts.org/en/factsheets.html.
2. Montavon, C.A.; Jones, I.P.; Staples, C.; Strachan, C.; Gutierrez, I.
Practical Issues In the Use of CFD For Modelling Wind Farms. Proc
European Wind Energy Conference, 2009.
3. British Wind Energy Association: England’s Regional Renewable
Energy Targets: Progress Report, 2009. http://www.bwea.com/pdf/
publications/RRETProgressReport.pdf.
4. Turbine Classifications: IEC-1400-1. http://windwire.blogspot.
com/2009/05/iec-classification-of-turbines.html.
5. BWEA: Calculations for Wind Energy Statistics – Emissions
Reductions. http://www.bwea.com/edu/calcs.html (accessed
16 July, 2009).
6. UK Renewable Energy Strategy (June 2008) HM Government
Department for Business Enterprise & Regulatory Reform (BERR).
7. BWEA UK Wind Energy Database.
8. Climate Change Act 2008 (Chapter 27).
9. Milborrow, D. Dispelling the Myths of Energy Payback Time. Wind
Stats, 1998 (Spring), Vol. 11, No. 2. Fluid dynamics results showing contours of velocity magnitude
10. Cabezon, D.; Iniesta, A.; Ferrer, E.; Marui, I. Comparing Linear and for Alaiz site. Fluid dynamics software from ANSYS was able to
Nonlinear Flow Models, Proceedings of EWEC 2006. accurately model this complex terrain.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 29


ENERGY: wind

Image © iStockphoto.com/chromatika
Second Wind
Advanced turbulence models lead
to optimized wind turbine spacing.
By Thomas Hahm, Fluid & Energy Engineering GmbH & Co. KG,
Hamburg, Germany

With the Global Wind Energy Council projecting that wind Atmospheric flow over the Bolund peninsula (Denmark) with wind
power installations will grow to 409 gigawatts in 2014, the direction from the escarpment on the west side. The red color
indicates high wind speeds, and blue indicates low speeds.
amount of land occupied by wind farms is becoming a
serious concern. The issue is already critical in Germany
because of dense population and the country’s leadership
in deploying wind power. Typically the goal is to generate
as much power as possible at the lowest cost from a
given wind farm site. But there are drawbacks to spacing
wind turbines too close together, including lower power
output and wake effects.
Fluid & Energy Engineering (F2E) GmbH & Co. KG is
pioneering the use of advanced turbulence models such
as large eddy simulation (LES) to increase the accuracy
with which turbines can be sited within a wind farm. The
end result will be the ability to generate more energy from
a given volume of land.
The wake of a turbine has significant effects — including
reduced power output and shorter turbine life — on any Wind field behind an ENERCON E-66 wind turbine with hub height of
65 meters. The blue color indicates high wind speeds, and white
turbine in its path. Extracting energy from incoming wind indicates low speeds.
causes a loss in the kinetic energy and velocity of the wind
in the wake of a turbine. This energy is recovered over dis-
tance as the wake exchanges energy with the surrounding
wind. Turbulence produced in the wake of turbines and
turbulence produced by terrain features such as forests or
hills can have a substantial impact on turbine life. The
effects of turbulence can vary with the flow angle and
inclination of the incoming wind. For example, one of the
most damaging scenarios is when one half of a rotor
experiences turbulence from an upwind turbine and the Meandering wake behind an ENERCON E-66 wind turbine. Blue
other half is exposed to undisturbed flow. represents high wind speed, and white represents low speed.

30 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


energy: Wind

In Germany, loan organizations and regulatory author- 0.45

ities require an assessment of every new wind farm project 0.4

to ensure that turbulence loads generated by terrain and 0.35

wind turbine wakes are within acceptable design limits. 0.3


Terrain and wind turbine wakes must be considered
0.25
together because they have additive effects on turbu-

TI[-]
0.2
lence. Typically, the approach to computing these loads is
0.15
to use standard empirical formulas to estimate terrain-
generated and wake-generated turbulence intensity and 0.1

to use blade momentum theory to estimate the loading on 0.05

the blades. 0
-150 -125 -100 -75 -50 -25 0 25 50 75 100 125 150
However, the accuracy of these empirical calculations Width of Wake [m]

can be limited because they do not incorporate the actual Simulation of local turbulence intensity (blue line) compared to
measurements (red dots) across width of wake
geometry of the rotors and terrain. The result is that the
calculations need to be heavily calibrated to produce
safe wind farm designs. These empirical models show
that wind turbines typically must be spaced at a distance Ultra Sonic 2-D LES Simulation 2-D
of about four rotor diameters from each other. Required 10
9
spacing can be even greater if terrain-generated 8
turbulence is a major factor. 7
Velocity [m/s]

Developers of wind farms are demanding more 6


5
accurate methods for calculating the turbulence intensity 4
generated by wakes and terrains so that they can produce 3
2
more energy from a given parcel of land. F2E has been 1
using computational fluid dynamics (CFD) over the last 0
100 120 140 160 180 200 220 240 260 280
decade to address this important issue. The company’s I [s]

engineers generally have had little difficulty in determining Measured (green line) and calculated (blue line) values for velocity
magnitude at hub height in center of wake over 180 seconds
the wind velocity in the wake of a turbine, but calculating
the turbulence intensity is much more challenging.
Conventional Reynolds-averaged Navier–Stokes the flowfield. Most recently, F2E engineers used ANSYS
(RANS) models reduce the computational time required to FLUENT software to model the full geometry of two
simulate turbulent flows by time-averaging the velocity ENERCON E-66 wind turbines with 66 meter rotors using
field, pressure, density and temperature over time. This the LES technique. The fluid dynamics results were
approach eliminates turbulence fluctuations and makes it validated with data collected in an actual wind farm using
possible to model turbulent flow in a reasonable time on ultrasonic anemometers. Fluid dynamics accurately
desktop computers. RANS methods are effective at pre- predicted the velocity of the incoming wind along with
dicting the overall and steady-state behavior of a wind turbulence intensity at the downwind turbine over the
farm; however, their accuracy suffers in modeling unsteady width of the wake.
turbulent flows that are typically found in wind farms. Fluid dynamics did a good job of predicting the varia-
Over the past three years, F2E has used the LES tion in the velocity over a period of 300 seconds, a figure
turbulence model with very positive results. It numerically that can be used to calculate turbulence intensity and the
resolves the larger turbulence scales and models the aerodynamic loads that turbulence creates. Quick hori-
smaller scales to provide accurate transient solutions of zontal shifts of the wake from one side to another are
detectable on a 25 second time scale. Another critical
factor in achieving this level of accuracy was the accurate
modeling of the incoming wind field and, particularly, the
variation in its wind direction and the wind shear.
Fluid dynamics and LES techniques from ANSYS can
be used to calibrate and greatly improve the accuracy of
empirical calculations that are currently used to calculate
blade loads under wake conditions. Calculation of blade
loads should make it possible to substantially increase
the amount of energy that can be generated by new
wind farms and to increase the safety and the lifespan of
Model of a wind turbine used for computational fluid
dynamics simulation wind turbines. n

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 31


Aerospace

Computer-generated image of the Terrafugia Transition in flight. The company unveiled a production prototype in 2010.

Fast Lane to Sky High


Fluid flow simulation software co-pilots design
of production prototype roadable aircraft.
By Gregor Cadman, Engineer, Terrafugia, Woburn, U.S.A.

Since the earliest days of the aviation industry, inventors available street-legal aircraft with hopes of beginning
and entrepreneurs — from motorcycle racers to homebuilt- manufacturing in 2011. Terrafugia — Latin for “escape
aircraft enthusiasts to the largest automakers — have from land” — used simulation tools from ANSYS to arrive
sought to develop the iconic mashup of future transpor- at a production prototype of its innovative Transition ®
tation technology known as the “flying car.” Some of their Roadable Aircraft. After earning global attention following
attempts did manage to test successfully and even reside successful test flights in early 2009, the new prototype
in the Smithsonian. Sporting names such as Autoplane, was revealed to the industry in 2010 at the annual
Aerobile and Airphibian, these machines were impressive Experimental Aircraft Association (EAA) AirVenture
for their time, but they never lived up to their mythological Oshkosh national airshow. It was a critical step in
pedigrees as conjured by science fiction authors and commercializing this one-of-a-kind vehicle.
filmmakers. The concept is a proven one. But due to the With a flight range of up to 490 miles and cruising
engineering challenges involved in combining a light- speed of 105 mph, the Transition can also drive up to
weight, aerodynamic aircraft with the stability needed for 65 mph on the road. It is capable of transforming from
long-distance driving, to date none of these vehicles, plane to car in less than 30 seconds. The sophisticated
which are more accurately called “road-ready” airplanes, design features foldable wings that span over 26 feet, a
have been successfully brought to market. rear-wheel-drive system for the road, and a rear pusher
Recently, however, the Boston-area startup Terrafugia propeller for flight. While the Transition is not designed to
revealed its production prototype for the first commercially replace anyone’s car, it is intended to drive at normal

Rear view showing pathlines passing by Terrafugia’s beta prototype design flying at 105 mph

32 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Aerospace

The Transition’s foldable wings allow the aircraft to be driven on the highway as well as parked in a typical residential garage.

highway speeds, so that owners can easily access their airflow tests to study the effects of the slightest design
local airports. This required Terrafugia’s engineering team change on overall performance. The team addressed
to simultaneously consider the aerodynamics of flying issues such as maximizing wing lift in the air while
and driving — activities that exert very different forces on simultaneously minimizing the effects of crosswinds
the vehicle. While physical tests in a wind tunnel helped to along the road. Without an ability to work in a virtual
validate the initial concept design, they were both time- environment, Terrafugia’s team would have had to
and cost-intensive. For this reason, Terrafugia engineers construct complicated physical models, modify or rebuild
turned to ANSYS FLUENT software to make and verify them, and conduct hours of real-world testing. Simulations
design modifications for the new production prototype, powered by fluid dynamics software from ANSYS enabled
working in a virtual simulation environment that saved rapid testing and verification of some modifications to the
time and money while also enabling engineers to assess a Transition design, based on the physical performance of
complex range of design considerations. the initial proof-of-concept vehicle.
Unlike a typical car or plane, the Transition has a host Wind tunnel tests revealed an adverse interaction
of extra components to consider when analyzing air flows between the vehicle’s front suspension and its canard,
around the vehicle, as wheels, propellers, foldable wings which serves as a wing while flying and a front bumper
and other shapes affect dynamic flows whether it is while driving. Rather than relying on additional physical
driving or flying. Using the modeling capabilities of the testing, Terrafugia used ANSYS technology to further
software, Terrafugia engineers conducted whole-vehicle explore this interaction as well as potential solutions. As
the work of the engineering team progressed, it became
clear that, while the canard configuration had initially been
integral to the design, it was undesirable from a number of
standpoints. With the Transition receiving classification as
a multipurpose passenger vehicle, the full-width bumper
requirement for passenger cars — the original reason for
the canard — was no longer applicable. Engineering
simulation software verified that lighter weight, better flight
characteristics and improved looks were all potential
benefits of a canard-free design.
Another challenging aerodynamic design aspect of the
Transition was attaining a wing stall speed — the speed at
Proof-of-concept design of the Transition showing pressure contours
on the vehicle surfaces. Since the Transition was classified as a multi-
which an aircraft stops flying — of under 52 mph, which is
purpose passenger vehicle, a bumper was not legally required, so a requirement for the light sport category of aircraft.
engineers conducted simulations to see if they could eliminate this
feature. ANSYS models confirmed that the canard was not needed, and
Since slower in-air speeds generally create safer flying
this feature has been eliminated in the production prototype design. conditions, it was important for the Terrafugia engineers

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 33


Aerospace

Production prototype design showing airflow pathlines over the vehicle body. The VBM plug-in to ANSYS FLUENT enabled
Terrafugia engineers to model the vehicle’s propeller under near-stall conditions, which helped to ensure the safety of the
aircraft while in flight.

to design the vehicle to operate at a low speed for safety the remainder of the body as needed to match the weight
and stability without stalling. Stall prediction can pose a and center-of-gravity requirements of the vehicle.
difficult problem, even for sophisticated CFD tools. The new design improves both the in-air and on-road
However, with close support from ANSYS experts, performance of the Transition as well as ensures that the
Terrafugia developed a detailed engineering approach vehicle lends itself to full-scale manufacturing. Terrafugia’s
necessary to obtain accurate predictions, including the team found that simulation software from ANSYS was
use of a virtual blade modeler (VBM) that plugged in to the critical in advancing them to the production prototype
fluid dynamics software and created additional capabil- stage with a high degree of confidence in their design.
ities to model the Transition’s propeller. After deploying The company credits ANSYS tools for helping to bring
the specialty VBM tool, the team reshaped the wing and their product to market so quickly. n

Windows®. Life without Walls ™.


HP recommends Windows 7.

Powering
with the
HP Z800 Workstation
with Intel® Xeon® processor

Visit www.hp.com/go/ansys for more information

© 2010 Hewlett-Packard Development Company, L.P. The Intel Logo, Xeon and Xeon Inside are trademarks of Intel Corporation
in the U.S. and/or other countries. Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation.

34 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


electronics

Image © iStockphoto.com/Cristian Andrei Matei


Successful Launch
Circuit and field tools combine to optimize satellite multiplexer
design, reducing time from 10 weeks to two days.
By ANSYS Advantage Staff

M1, j
The traditional approach to designing M1, i
R1 R2
multichannel microwave switches 1F 1F 1F 1F 1F 1F

and multiplexers used in satellite i1 i2 ii ij iN-1 iN

communications is to build them 1H 1/2H 1/2H 1/2H 1/2H 1/2H 1/2H 1/2H 1/2H 1H
empirically channel by channel, which
takes about one week per channel, or (1) (2) (i) (j) (N-1) (N)

10 weeks for a 10-channel device. M2, i Mj, N-1


M1,2 M2, j MN-1, N
A team of engineers at COM DEV,
a space hardware manufacturer, has
RLC model of one channel of multiplexer
developed a new multiplexer design
method that intelligently links the frequencies to maximize the output microwave multiplexers. A
HFSS 3-D full-wave finite element communications capacity of the typical 10-channel output multiplexer
electromagnetic (EM) simulator from satellite. This creates the need for contains 140 design variables. Full-
ANSYS with a faster but less accurate multiplexers, which separate the wave EM simulation can accurately
coupled RLC circuit model. Six or received signal into frequencies used model the performance of a multi-
seven iterations of this process, which by each transponder. At the same plexer; however, optimizing the perfor-
take about two hours each, are time, the multiplexers serve as micro- mance of all 140 design parameters
sufficient to optimize the design, a wave filters with little attenuation in with dielectric material would require
total of about two days. their passbands and steep skirts at thousands of simulation runs. The
A communications satellite acts the edges of their bands. Additional traditional approach has long been to
as a repeater high above the globe. multiplexers are used to combine use physical experiments to build and
RF signals from the ground station the output signals of the transponders test multiplexer devices one channel
travel toward the satellite, where they into a single RF signal for the return at a time. This approach is time
are received, converted, amplified trip to earth. consuming and highly dependent on
and transmitted back to Earth. COM DEV technology is used on the design optimization skills of
A transponder is the hardware used more than 700 spacecraft, including technicians and engineers.
to take in signals on one set of 80 percent of all commercial communi- COM DEV developed a more
frequencies and then resend to cations satellites ever launched. efficient and robust simulation method
the ground on other frequencies. Most fixed-service communications that uses space-mapping simulation
Multiple transponders use different satellites use COM DEV input and to link HFSS with a fast but not

Short
circuit Tuning screw

T-J T-J T-J

Common
port

Channel 1 Channel 2 Channel N


Dielectric Dielectric
Channel inputs support resonator

Manifold multiplexer model Channel design example of a five-pole DR filter

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 35


Electronics

very accurate RLC circuit model, Multiplexer Design Example -10


combining the best features of both An example shows how this

|S11| in (dB)
tools. The space-mapping concept method works in practice: a -20

is based on mapping between 10-channel manifold-coupled output -30


optimizable design parameters of multiplexer in the frequency band of
t h e f i n e m o d e l a n d p re c i s e l y 3.5 GHz to 4.25 GHz. The multiplexer -40

corresponding parameters of the comprises a number of channels,


-50
coarse model. This approach differs each consisting of a narrow bandpass 3.57 3.6 3.65 3.7 3.75
Frequency (GHz)
from earlier space-mapping algo- filter connected to a waveguide
rithms as it establishes implicit rather manifold. Eight channels have a Ideal (solid line) and HFSS simulated
(dashed line) return loss (S11) response
than explicit mapping between the bandwidth of 1.5 percent, and the of channel 1 after seven iterations
fine and coarse models. In each remaining two have a bandwidth of
0
iteration, the implicit space-mapping 0.8 percent. The equivalent circuit
-10
algorithm extracts selected pre- of one channel of this filter is used as
-20
assigned parameters to match the a coarse model. As a rule of thumb,

|S21| in (dB)
coarse model with the fine model. RLC models are valid when electrical -30

With these pre-assigned parameters length is less than one-tenth of the -40

fixed, the calibrated coarse model wavelength (λ). The wavelength at -50
is re-optimized, and the optimized 3 GHz is approximately 100 milli- -60
design parameters are then assigned meters, so the transmission line and
-70
to the fine model. This process radiation effects not captured with 3.57 3.6 3.65 3.7 3.75
is repeated until the fine model an RLC model are important, and an Frequency (GHz)

response is sufficiently close to S-parameter model is required for a Ideal responses (solid line) and HFSS
simulated (dashed line) insertion loss
the target response. All channels of complete description of the multi- (S21) response of channel 1 after
the multiplexer can be optimized plexer behavior. The geometric seven iterations

simultaneously. dimensions of the manifold and the algorithm to the overall circuit model
channels were optimized by applying of each channel to optimize the
the space-mapping optimization channel dimensions.

Rapid Design of RF and Microwave Components


Rapid design of an RF and microwave component can However, it is still possible to rapidly design,
be easily accomplished when HFSS and DesignerRF optimize or tune these very complex components. For a
products from ANSYS are used together. Employing full RF and microwave component, the simpler basic
these two tools in combination allows a user to design, elements are created, parameterized and rapidly
optimize or tune a device with the accuracy of HFSS simulated in HFSS. The full parametric solution of each
at a speed that rivals the time that it takes to perform building block is then stored in a design element
a circuit simulation. repository. Using any combination of elements from this
RF and microwave components often comprise repository, the full RF and microwave component can
basic building blocks or repeating elements. These be assembled or synthesized inside DesignerRF. Since
blocks are usually simple and are quickly and easily all the individual elements are parameterized and
simulated using HFSS. When these basic blocks are electromagnetically characterized, the final assembly
combined into a more complex structure, the electro- can be rapidly tuned using the DesignerRF tool.
magnetic simulation time can increase dramatically.
Markus Kopp, Product Manager, Electronics, ANSYS, Inc.

Input HFSS element HFSS iris element DesignerRF circuit model with three iris and two input elements

1:1 2:1 2:1 2:1 2:1 2:1

Port 1 Port 2
Filter response

A multi-pole waveguide cavity filter consisting of four cavities and five irises using three iris and two input elements can quickly be created in
DesignerRF using iris and input elements. For each of the individual elements, the dimensions of the iris and waveguide length can be
parameterized and solved in HFSS, the five elements can be combined in DesignerRF, and a given filter response can be determined by
changing the available dimensional variations.

36 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


electronics

0 0

-20 -20
IL (dB)

IL (dB)
-40 -40

-60 -60

-80 -80
3.55 3.75 3.95 4.15 4.35 4.5 3.55 3.75 3.95 4.15 4.35 4.5
Frequency (GHz) Frequency (GHz)

Ideal response of the multiplexer (solid line) Measured response of the multiplexer prototype
compared to the HFSS simulated response
(dashed line) of the final design

An accurate multiplexer model magnetic saturation. The next step the optimal channel design
was obtained by replacing each was defining boundary conditions that parameters over the multiplexing
channel with the corresponding specify the field behavior on the sur- frequency band of interest. The
parameters sweep provided by HFSS faces of the solution domain and resulting multiplexer model takes
at the optimal channel dimensions. object interfaces, such as lumped RLC into account the effects of channel
HFSS technology can accurately for tuning elements and impedance dispersion and spurious modes.
model multiplexers and other passive for thin film resistors. Ports were Finally, the manifold parameters were
microwave components down to fine defined where energy enters and exits re-optimized to meet the required
details, such as tuning screws and the model. specifications. The optimization
probes. HFSS employs the finite To calculate the S-matrix of a variables included manifold spacing
element method on small unstructured structure with ports, HFSS auto- between channels and lengths of
mesh elements (where needed) and matically defined a geometrically waveguides connecting the channels
large elements (where small elements conforming tetrahedral mesh. It then to the manifold. By optimizing the
are not needed) to reduce processing computed the modes on each port of new multiplexer model, the manifold
time without sacrificing accuracy. the structure that are supported by a design parameters were adjusted to
Adaptive meshing refines the mesh transmission line having the same compensate for channel dispersion
automatically where field accuracy cross section as the port. HFSS and spurious modes.
needs to be improved. computed the full electromagnetic
The engineer began the modeling field pattern inside the structure, Optimization Achieves
process by importing the geometry of calculating all modes and all ports Close-to-Ideal Design
the initial design from a computer- simultaneously for the 3-D field COM DEV has used this approach
aided design (CAD) file. The engineer solution. The next step was computing to design a large number of multi-
then defined the electrical properties the generalized S-matrix from the plexers and other passive microwave
of the materials, such as permittivity amount of reflection and transmission components. The company has
and dielectric loss tangent, perme- that occurs. The resulting S-matrix dramatically reduced the amount of
ability and magnetic loss tangent, allowed the magnitude of transmitted time and money required to design
bulk electrical conductivity, and and reflected signals to be computed each component. The process has
directly from a given set of made it possible to bring new
input signals, reducing the products to market at a much faster
full 3-D electromagnetic rate. At the same time, the quality of
behavior of a structure to the designs has been improved, since
a set of high-frequency the new method is repeatable and
circuit parameters. verifiable and does not rely on the
The coupling matrix skills of the individual engineer. n
of the coarse model Reference
was then fitted with Mostafa, I.; Smith, D.; Panariello, A.; Wang, Y.;
th e cor res pon di n g Yu, M. EM-Based Design of Large-Scale
S-parameter sweep Dielectric-Resonator Filters and Multiplexers
by Space Mapping. IEEE Transactions on
block computed by Microwave Theory and Techniques, 2004,
10-channel output multiplexer the EM simulator at Vol. 52, No. 1.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 37


Scientific Equipment

Glass Jaw 5mm

Simulation helps to solve collimator jaw


design problem in the Large Hadron Collider.
By Alessandro Bertarelli and Alessandro Dallocchio, Engineering Department, Mechanical
and Material Engineering Group, European Organization for Nuclear Research, Geneva, Switzerland Collimator cross section

The Large Hadron Collider (LHC) at the European Sited near Geneva, Switzerland, the LHC is being
Organization for Nuclear Research (Conseil Européen pour exploited by physicists to study the smallest known
la Recherche Nucléaire, or CERN) is the world’s largest particles that are building blocks of all things. The team
and most powerful particle accelerator. Two beams of uses the LHC to re-create conditions that existed just after
subatomic particles called hadrons — either protons or the big bang by colliding two beams head-on at very high
lead ions — travel in opposite directions inside the circular energy. The target of study is the Higgs boson, a theorized
accelerator, gaining energy with every lap. The LHC’s but not yet discovered particle that may help explain the
collimators scrape away particles that have gone slightly masses of various subatomic particles. The LHC is being
off track to prevent damage to the highly sensitive super- used to seek out dark matter that is believed to make up
conducting magnets. But the collimators need to be able 96 percent of the mass of the universe. Scientists hope to
to withstand an error that might pound them with a discover the differences between matter and antimatter
substantial fraction of the beam itself — considering that to help in understanding why so little antimatter is left in
the beam could melt almost 1 ton of copper. the universe even though the big bang is believed to have
When researchers tested the LHC collimator prototype produced equal amounts of matter and antimatter.
with several shots at different beam intensities, they The LHC uses a stored energy of 360 MJ per beam.
discovered that the carbon–carbon collimator jaw nicely This is two to three orders of magnitude above what other
survived the impact, but its metal support suffered a proton colliders can handle. CERN’s collider is contained
permanent deflection strong enough to put the collimator in an underground circular tunnel, 50 meters to 175 meters
out of action. CERN staff immediately set to work to under- deep, with a circumference of 27 kilometers. The collider
stand what caused the problem and how to correct it. tunnel contains two adjacent parallel beam pipes, and
each contains a proton beam that rotates in opposite
directions around the ring. The parallel beams intersect
Cooling pipes – internal
pressure 15 bar at four points, forcing them to collide with each other.
Some 1,232 dipole magnets keep the beams on a circular
path, and an additional 392 quadrupole magnets focus
the beams to increase the chances of collision at the inter-
section points.
3-D contact elements
The most critical elements of the collimators are the
with friction jaws, which are made of carbon–carbon composites,
which encircle the beam and are designed to block any
SOLID45 – eight-
node brick element stray particles that separate from the beam. The jaws are
with initial strain
to simulate spring supported by a brazed sandwich structure encompassing
preload (three bar)
Finite element model of collimator the main support bar, cooling pipes and interface plate.

38 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Scientific Equipment

The collimator is at risk when the beam is moved from


one accelerator line to another. This needs to be done
frequently because the beam is accelerated to maximum
energy by moving it through a family of progressively more Flexural vibrations

powerful accelerators. Kicker magnets are used to knock


the beam out of its current line and inject it into the new
line. If there is an error in the configuration of the kicker
magnets, the beam might be steered too much
in one direction. The collimator must be able to with-
stand accidents such as an injection error that may cause
3 . 2 x 1 0 1 3 p r o t o n s e a c h w i t h a n e n e rg y o f
450 gigaelectronvolt (GeV) to hit the collimator jaws. An Temperature distribution in
electron volt is the kinetic energy gained by an electron collimator, where blue is minimum
and red is maximum
when it accelerates through an electric potential difference
of 1 volt.
To validate the collimator design, engineers carried finite element analysis software provides multiphysics
out tests using another CERN accelerator on a fully capabilities that integrate thermal and mechanical
operational prototype of a collimator. The jaws were analysis. In this application, integration increases
submitted to a series of impacts at 450 GeV over 7.2 μs. accuracy and reduces the amount of time required to
Measurements performed on jaw assemblies and metal prepare the model. ANSYS Mechanical also offers compre-
supports revealed a permanent deformation of the metal hensive element technology, an extensive library of
support of over 300 μm. CERN researchers theorized material models, auto contact detection for assemblies
that this deformation was caused by thermally induced and powerful solver capabilities.
vibrations due to very fast heating. Even when a structure The collimation block was modeled as a rectangular
is free to expand, when the heating process is faster than beam, simply supported at its edges. The energy distri-
the typical stress relaxation time, material inertia prevents bution applied to the collimator was determined with
free thermal expansion, causing stress waves. Simulation FLUKA, a particle physics Monte Carlo simulation
was clearly required to better understand the problem. package. The beam energy was introduced to the finite
Applications with such great mechanical complexity element model in the form of a 3-D table using the HGEN
are typically addressed with explicit dynamics codes. command. Transient thermal simulation was used to
But for this case, explicit analysis would have been very calculate the temperature distribution as a function of
awkward because of the need for multiphysics integration time. Researchers performed structural dynamic analysis
and the complexity of the mechanical structure, which by applying the temperature distribution as nodal loading
includes different materials that are held together with at different time steps. Elastoplastic analysis used the
clamps. Since the stresses are well below the elastic multilinear kinematic hardening model for metallic
modulus, there are no shock waves that would require an components. The team developed special algorithms to
explicit dynamics code. The problem could instead be apply a temperature distribution changing in time and
simulated using an implicit finite element model following space at different substeps of the elastic–plastic analysis.
the rules of thermo-elasticity. ANSYS Mechanical implicit The integration time step of 0.1 μs was based on the
0

-0.05
Transverse deformation (mm)

-0.1

-0.15

-0.2

-0.25

-0.3
100 200 300 400 500 600 700 800 900 1000
CFC assy CFC assy Graphite bloc Graphite support

Jaw Longitudinal Position (mm)

Residual displacement based on physical measurements (left) and simulation (right) matched very closely.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 39


Scientific Equipment

2.50E-04
Measured displacement - 1 batch Measured displacement - 2 batches

2.25E-04 FEM - scaled to 1 batch 1dof - no damping


1dof damped

2.00E-04

1.75E-04
Displacement [m]

1.50E-04

1.25E-04

1.00e-04

7.50E-05

5.00E-05

2.50E-05

0.00E+00
0.E+00 1.E-02 2.E-02 3.E-02 4.E-02 5.E-02 6.E-02 7.E-02
Time [s]
Simulation predictions correlate well with measured deflection.

preliminary analytical estimation to avoid numerical deflection of 350 μm predicted by the simulation closely
damping. Finally, static analysis evaluated residual plastic matched the measured value of 300 μm. The dynamic
deformation. response predicted by the simulation also correlated well
The simulation results included the temperature rise with laser Doppler vibrometer measurements. The
throughout the structure, making it possible to use simple simulations are higher in magnitude than the physical
formulas to predict plasticization. The maximum stress is measurements because the simulations do not take
well above the proportional limit of copper, so the thermal damping into account.
shock generated plastic strains, as theorized by the The study team, made up of senior and junior
researchers. The largest residual plastic strains of 0.12 engineers, fellows, and students, used the validated
percent were seen in the 3–mm thick copper plate. These simulation model as the primary tool to solve the problem.
strains are eccentric with respect to the neutral axis CERN researchers evaluated a number of different
of the metal support, leading to permanent deflection geometries and materials to determine their impact on the
away from the beam axis. simulation of the support structure. Using these results,
The simulation results matched the physical measure- they decided to modify the jaw assembly series design by
ments remarkably closely. When comparing the physical changing the thin plate material from OFE-copper to the
measurements in the line graph to the simulation higher yield strength Glidcop® material. Glidcop is a family
predictions in the contour plot, the transverse residual of copper-based metal matrix composite alloys mixed
primarily with aluminum oxide ceramic particles. The
addition of small amounts of aluminum oxide greatly
increases the copper’s resistance to thermal softening
and enhances elevated temperature strength.
An updated model of the series jaw assembly
including Cu-Ni pipes and Glidcop support beam and thin
plates showed that deflection was reduced from 300 μm
to 16 μm, which was within acceptable limits. This
figure was confirmed by experimental tests carried out
on a second prototype of the collimator. The use of
simulation in this application made it possible to rapidly
diagnose the problem and develop an acceptable solution
while eliminating the need to build additional proto-
types, excepting the final design. ANSYS Mechanical
software played a key role by providing the full range of
physics capabilities needed to accurately simulate a very
Deflection was substantially reduced in the new design. complex problem. n

40 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


ANALYSIS TOOLS

Designing Batteries
for Electric Vehicles
Numerical simulation can be used to accelerate battery
development and address safety concerns.
By Xiao Hu, Lead Engineer, ANSYS, Inc.

X= X=

The lithium-ion battery is a preferred candidate as a 0


ᵟ n ᵟ s ᵟ p L

source of power for the hybrid electric vehicle (HEV) and


electric vehicle (EV) because of high energy density, high
voltage, low self-discharge rate and good stability. HEV
Current Collector

Current Collector
and EV applications require very large lithium-ion
batteries, but, during high power extraction required to
drive a vehicle, these large batteries may experience a
significant temperature increase — which can lead to
safety concerns. A properly designed thermal manage-
Negative Separator Positive
ment system is crucial to prevent overheating and uneven
Electrode Electrode
heating across a large battery pack, which can lead to
degradation, mismatch in cell capacity and thermal Schematic of a lithium-ion cell sandwich consisting of
composite negative and positive electrodes and separator
runaway. Design of the thermal management system
requires knowledge of the cooling system as well as the rate of heat generation varies substantially with time over
amount of heat that will be generated by cells within the course of charging and discharging. Heat can be
the battery pack. generated from internal losses of Joule heating and local
Simulation can assist in thermal design at both the electrode overpotentials, the entropy of cell reaction,
cell level (a single battery cell) and the system level mixing and side reactions. When only the most important
(a battery module or a complete battery pack). effects of Joule heating and local electrode overpotentials
On the battery cell level, the focus is on detailed heat are considered, heat generation can be expressed by
generation and temperature distribution within a battery open-circuit potential and the potential difference between
cell — information mainly used by manufacturers and positive and negative electrodes. By using models to
battery researchers. Experimental data reveals that the predict potential and current density distribution on the
electrodes of a lithium-ion battery as a function of
discharge time, the results can be used to calculate
the temperature distributions of the lithium-ion battery.
These temperature distributions can then be used within
ANSYS CFD software to examine the effect of the
configuration of the electrodes — the aspect ratio of
the electrodes and the placing of current collecting tabs
— as well as the discharge rates on the thermal behavior
of the battery.
While this type of model provides detailed information
about temperature and current density distribution, it
requires experimental testing data as input. This model
cannot predict the impact of design changes on battery
thermal performance without conducting testing again.
However, a physics-based electrochemistry model can
Current density distribution Temperature distribution be used to investigate the impact of battery design
within a battery cell within a battery cell
parameters on battery performance. The electrochemistry
Temperature distributions from this fluid dynamics simulation were shown model includes geometry parameters, properties and
to be in good agreement with those from the experimental measurement.

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 41


ANALYSIS TOOLS

a different set of simulation goals. Engineers using


computational fluid dynamics (CFD) for battery thermal
management are interested in maintaining the
desired temperature range, reducing pressure drop and
maintaining temperature uniformity. Fluid dynamics has
been widely used to predict flow and heat transfer in many
industries, and it applies well to battery thermal manage-
ment. By using the ANSYS Workbench platform with
ANSYS CFD software, a complete battery thermal–fluid
Automotive battery module with 16 cells dynamics analysis, including optimization, can be done
entirely within the same environment.
While fluid dynamics simulation can give detailed
thermal information about battery thermal management
systems, it is time consuming to perform many transient
simulations under different drive cycles. Model order
reduction techniques exist for extracting a model from
CFD results, and the extract model, called Foster network
model, gives the same solution as that from the full
CFD model — but it runs much faster compared with
Fluid flow solution for the 16-battery cell module CFD. The model order reduction process is handled auto-
A complete battery thermal–fluid dynamics analysis including
matically by Simplorer software using CFD results as
optimization can be done using ANSYS FLUENT software inputs. This process opens the door for simulations that
entirely within the ANSYS Workbench environment. would otherwise have been impractical, such as battery
thermal control system analysis.
temperature, with the last being the most important. A For electrical engineers, the primary concern is the
physics-based model also can provide inputs that would electric performance of the battery rather than the thermal
otherwise need experimental testing to obtain. The most
20
famous physics-based model originally was proposed by 4.3

Professor John Newman from UC Berkeley. Such a model 15


4.1
has been implemented in Simplorer software from ANSYS 10
to allow determination of many important factors in 3.9
5
battery design.

Cell Potential (V)


Current (A)

3.7
Charge and discharge cycle results can be plotted 0

along with concentration profiles during discharge. From -5


3.5

this information, optimization of the initial composition of 3.3


-10
electrolytes in the cell can be performed. When this study
3.1
was performed, results showed that a higher initial -15

concentration of electrolytes leads to as somewhat lower -20 2.9


0 1 2 3 4 5
conductivity in the separator but a much higher conduc- Time (hr)
tivity in the composite cathode, where this is extremely
Battery charge and discharge cycle results from John
important. By plotting concentration profiles under Newman’s electrochemistry model
different temperatures, battery designers can determine
2600
when the limiting current occurs to specify the temp-
erature range that the system needs to maintain to avoid
2400
reaching limiting current. In addition, battery runtime is a
strong function of temperature, and battery runtime is 2200
longer with higher operating temperature. This can be
c(mol/m^3)

confirmed from a physics-based electrochemistry model, 2000


but it presents another problem: Higher temperature
brings safety concerns, shortens battery life, and becomes 1800
t = 1 min

another optimization issue in battery design. t = 5 min

t = 10 min
System-level design engineers working at the module 1600
t = 40 min
or pack level have a different set of requirements. Typically,
1400
these engineers cannot afford to simulate as many details 0 0.2 0.4 0.6 0.8 1

as engineers working at the cell level can; they also have X


Concentration profiles during galvanostatic discharge

42 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


ANALYSIS TOOLS

299 performance. An accurate and yet simple-to-use thermal


ANSYS FLUENT Cell 3 model that couples with a battery electric circuit model is
ANSYS FLUENT Cell 10 required. This can be accomplished through VHDL-AMS,
298
ANSYS FLUENT Cell 15 an IEEE ® standard hardware simulation language
Cell Average Temperature (K)

Foster Network Cell 3 supported by Simplorer.


Foster Network Cell 10 Designing batteries for HEV/EV involves many
297
Foster Network Cell 15 challenges, and designers focusing on different aspects
have different requirements. The benefit with ANSYS
software is that it offers models ranging from cell-level
296
electrochemistry to system-level thermal management. n

295
Reference
0 500 1000 1500 2000 Hu, X.; Lin, S.; Stanton, S.; Lian, W. A Novel Thermal Model for HEV/EV
Time (Sec) Battery Modeling Based on CFD Calculation, Proceedings of IEEE Energy
Conversion Congress and Expo, Atlanta, U.S.A., September 12–16, 2010.
Comparison of CFD results with Foster network results. In Hu, X.; Lin, S.; Stanton, S.; Lian, W. A State Space Thermal Model for HEV/
this case, the Foster network model took 20 seconds to run,
EV Battery Modeling. SAE, 2011, 01-1364.
whereas traditional CFD required a few hours.
Fuller, T.F.; Doyle, M.; Newman, J. Simulation and Optimization of the Dual
Lithium Ion Insertion Cell. Journal of Electrochem. Soc., 1994, vol. 141,
pp. 1–110.

5RYITNOW
7ISITWWWMICROSOFTCOMHPC

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 43


Academic

Microbubbles Keep
Green Energy Blooming

Image © iStockphoto.com/kamil.
Algae-derived biofuel production gets
a strong pulse from flow simulation.
By William B. Zimmerman, Professor of Chemical and Biological Engineering, University of Sheffield, U.K.

The world faces a shortage of fossil To reduce these costs, one of the
fuels; at the same time, efforts are developing applications being stud-
being made to reduce carbon dioxide ied is to introduce CO2-rich fossil fuel
(CO2) emissions. Production of bio- exhaust gases into the bioreactor as
fuels from certain kinds of algae offers microbubbles. Among the groups
a potential solution to both problems. tackling this problem is a research
Like other photosynthetic organisms, Left: A conventional gas diffuser produces team from the chemical engineering
relatively large bubbles. Right: A gas diffuser
algae capture CO2 and sunlight and with the fluid oscillator shows tremendous department at the University of
then convert the elements to oxygen reduction in bubble size. Sheffield in the U.K. The group
and the energy required for their life has made a major step forward in
cycles. The algae biomass itself includes proteins, carbo- reducing the cost of producing algal biofuels from
hydrates and fatty acids. It is these fatty acids — or natural exhaust gases by creating microbubbles that are about
oils — that offer great potential as a renewable feedstock 20 microns in diameter. The major advantage of smaller
for refineries, where they can be transformed into fuel to bubbles is that they have a higher surface-area-to-
power gasoline or diesel engines. volume ratio. Because these 20 micron microbubbles are
However, most processes to manufacture algae- 50 times smaller than conventional 1 millimeter fine
derived biofuels are too expensive to be commercially bubbles, they have a correspondingly higher surface area
viable — since the lowest production prices in the United per unit volume. Thus, they provide 50 times greater mass
States are currently in the $20 to $30 per gallon range. transfer rates, which could potentially translate into a
Two of the largest costs involved are the capital expenses tenfold higher yield in algal biofuel production [1].
required to build the bioreactors (producing algae on a The key to producing bubbles this small is a unique
large scale) and the energy required to operate them. oscillator design developed using ANSYS FLUENT fluid
dynamics software (part of a fluid dynamics Academic
Product bundle from ANSYS) to evaluate the initial
Flow from high-pressure port X2 design concept and perform parametric studies to select
to low-pressure port X1
appropriate nozzle width and flow rates that provide the
right operating conditions. The traditional method of
producing bubbles for reactors has been to force the gas

Air supply
Feedback
loop

Fluidic
Bubbles diverter
Growing A amplifier B
blown
small air away
bubbles
Air bleeding
to atmosphere
Water tank

Water
flow
Cover plate
Air flow pulse pulse
shown
transparent

Schematic representation of fluidic oscillator Fluid oscillator driving microbubble generation system

44 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Academic

With the noted 50-fold increase in mass transfer rate


afforded by this oscillator design, CO2 dispersal is accel-
erated in the bioreactor and should enhance algae growth
rate by a factor of 10. The microbubbles more efficiently
strip the oxygen that algae emit during photosynthesis,
thus permitting much higher algae densities in the water.
These improvements should substantially reduce the
capital and operating expense required for a given volume
of biofuel production, as more algae can be grown more
quickly in a smaller amount of space.
Top view of velocity magnitude contours on midplane of oscillator Beyond displaying a supercharged growth rate,
showing supplied gas flow directed into the upper outlet channel and
the experiments showed an 18 percent reduction in the
no flow through the two teardrop-shaped control nozzles
energy required for bubble production compared to
under pressure through a lattice of very small holes. When conventional fine bubbles. Researchers attribute this,
finally separated from the nozzle, however, the bubble has in part, to reduced friction losses due to the oscillatory
a diameter that is often many times larger than the hole, flow. The team thus predicts that engineering simulation
because its separation is controlled by the surface tension will play an even greater role in the coming design of a
of the water. The key idea of the Sheffield team’s micro- commercial-scale bioreactor, when it will become critical
bubble generation method was to limit bubble growth to optimize all components of the design to minimize
time using an oscillator to supply gas to the bubble capital expenses. n
formation holes. Growth is terminated at the end of each
pulse of gas. Bubbles are then removed from the nozzle References
using a pulse of water that alternates with the gas pulses [1] Zimmerman, W.B.; Zandi, M.; Bandalusena, H.C.H.; Tesař, V.; Gilmour,
so the growth of new bubbles can begin in the next oscil- D.J.; Ying, K. Pilot Scales Studies of Microbubble Mediated Airlift Loop
Bioreactor Growth of Microalgae Dunaliella Salina. Applied Energy,
lation period. Using this method, no bubble can reach the submitted in 2010.
size typically formed from steady gas blowing. [2] Tesař, V.; Bandalusena, H.C.H. Bistable Diverter Valve in Microfluidics.
In the oscillator design [2,3], steady gas flow is Experiments in Fluids, 2010; doi:10.1007/s00348-010-0983-0.
supplied to the oscillator inlet. The main flow stream then [3] Zimmerman, W.B.; Hewakandamby, B.N.; Tesař, V.; Bandulasena,
exits one of the two outlet channels depending on the H.C.H.; Omotowa, O.A. On the Design and Simulation of an Airlift
action applied by flow through a bidirectional control loop. Loop Bioreactor with Microbubble Generation by Fluidic Oscillation.
Food and Bioproducts Processing, 2009; Vol. 97, No. 3, pp. 215–227.
The flow exits through the upper outlet channel, causing
a decrease in pressure at the upper control nozzle. The
upper control nozzle then draws gas through the feedback
loop from the lower control nozzle, where the pressure is
higher. After the flow in the feedback loop tube gains
momentum, the flow exiting the upper control nozzle
causes the main jet — because of inducement (or amplifi-
cation) effects — to divert to the lower outlet channel.
After a delay needed for the feedback flow to gain
momentum in the opposite direction, the main jet is
diverted back to the upper outlet, resulting in a periodic 3-D view of gas flow pathlines released from a plane cut in the z
switching process. The frequency of the oscillation is direction and colored by velocity magnitude. The cusped splinter nose
controlled primarily by the length of the feedback loop at the junction of the outlet channels generates an internal feedback
loop that initially helps to direct the main flow to the upper outlet.
and the gas supply flow rate.
Sheffield’s team performed simulation of an early
oscillator prototype with ANSYS tools. Researchers
meshed the 3-D geometry with 70,000 tetrahedral cells
and used the renormalization group (RNG) k-ε model to
handle turbulence at lower Reynolds numbers. The results
matched the performance parameters of the early proto-
type, so the team proceeded to use fluid flow simulation to
improve the device’s controllability, including virtual testing
of geometries with different control nozzle widths and
different input flow boundary conditions. The researchers
Pathlines later in the cycle showing that flow through the control nozzles
compared the results for the different design alternatives has overcome internal feedback and the main jet is being switched to
and then selected the final proof-of-concept design. the lower outlet channel

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 45


Academic

Reforming a Fuel Cell


Modeling Process
Coupling flow simulation with complex chemistry tools brings
a united front to analyzing leading-edge energy systems.
By Robert J. Kee, George R. Brown Distinguished Professor of Engineering, Colorado School of Mines, Golden, U.S.A.

Since its basic principle was first temperatures offer the


demonstrated in the early nineteenth side benefit of using the
century, fuel cell technology has exhausted heat energy.
evolved into many different variations. With support from the
The underlying mechanism common U.S. Office of Naval
to all fuel cells is conversion of Research, a team from
chemical energy into electricity by the Colorado School of 850˚C
means of reforming the fuel into Mines (CSM) has been
hydrogen along with the subsequent working with ANSYS
electrochemical oxidation of hydrogen simulation tools to
into water. Depending on the type of model the chemistry, 750˚C

fuel cell and its application, the fuel electrochemistry and Outer shell temperature
Tube-wall temperature and
shell-interior streamlines
can be lighter hydrocarbons — such fluid mechanics of an
Results from the simulation of a shell-and-tube reformer
as natural gas, propane or methanol SOFC stack system. simulation for the catalytic partial oxidation of propane in which
— or heavier liquids, like diesel or Such a system — for an ANSYS FLUENT model of three-dimensional fluid flow and
jet fuel. The key advantages of fuel example, an auxiliary heat transfer is coupled with a CHEMKIN-based plug-flow
model. Shown are the temperatures on the outside of
cells over systems that burn fossil power unit (APU) used containment shell (left) and catalyst tubes (right).
fuels include fewer moving parts and by a Navy vessel —
overall reduced pollutant emissions. commonly comprises a the air flow surrounding the tubes is
Some of the challenges in developing shell-and-tube design that includes c o m p l e x b u t d o e s n o t i n v o l v e
fuel cells for more widespread use inter nal or exter nal reformers, chemical complexity.
are the high cost of catalyst or other depending on the fuel. The endo- To evaluate the full anode-
fabrication materials, the difficulty of thermic steam-reforming operation supported SOFC stack configuration
hydrogen storage, and very complex is supported by circulating the — in which the anode side is the
chemistry. e x h a u s t e d h e a t e n e r g y f r o m tube side and the cathode side is
Solid oxide fuel cells (SOFCs) in the exothermic electrochemical the shell side — the CSM team
particular have been the subject of oxidation within the fuel cell. The needed to couple the complex
much research in recent decades: complex catalytic chemistry is chemistry with the three-dimensional
They have the ability to reform many confined within the tubes, while the fluid mechanics. On the shell side,
different fuels, and their high operating three-dimensional fluid mechanics of researchers considered the fluid flow

Waste heat inlet


Reforming agent
Catalyst tubes

Fuel
vapor

Reformate

Swirl mixer Waste heat exhaust

Conceptual configuration for a shell-and-tube reformer Configuration of a 36-tube


anode-supported SOFC stack

46 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Academic

Cathode
Electrolyte
Anode O2 mole fraction
Tube temperature
810˚C 21.0%

530˚C 13.5%

Illustration of a single anode-supported SOFC tube. Electric Results from modeling a 66-tube SOFC stack operating on a
current is generated on the inside of the tubes (anode) and mixture of hydrogen, carbon monoxide and methane, which
is discharged on the tube exterior (cathode). The balloon are produced by reforming hexadecane. The right-side tubes
shows the essential microscale electrochemical phenomena show gas-phase oxygen mole fraction on the tube surfaces.
in the porous composite electrode layers. The left-side tubes show tube surface temperatures.

and heat transfer, including thermal function (UDF) capability, the research because it is important for all tubes to
radiation among all tubes and the team automated the process of deliver similar performance. The
containment shell. For this task, the averaging three-dimensional coupled model is useful for
team chose ANSYS FLUENT software temperature and oxygen mole fraction investigating the effects of design
(a component of the ANSYS data from the shell-side simulation considerations, such as tube packing
Academic Research CFD product and mapping it onto the tube-side and air-flow alternatives, on the
bundle) to model the complex, but band mesh. Additionally, the UDF overall performance of the stack.
nonreactive, fluid mechanics. directed the tube-side chemistry However, since the tube interior
On the tube interior (anode) side, model to supply heat flux and oxygen geometries were not fully resolved in
the simulation of chemical kinetics for mass flux boundary conditions from this research due to their complex
reforming practical military logistics each of the tubes back to the shell- microstructure, extensions to the
fuels — diesel in the case of an side fluid flow model. work under consideration include
APU — demands hundreds of surface Results of the simulations performing detailed three-dimensional
reactions and thousands of gas- revealed that there can be significant ANSYS FLUENT simulations of the
phase reactions. One-dimensional temperature variations between charge transport through the porous
chemistry tools such as CHEMKIN™ different tubes. The tubes were anode material on representative tube
or CANTERA can handle the reaction generally cooler at the bottom, which volume sections to calculate the
kinetics and charge transfer, as was caused by a combination of effective electrical conductivity. This
long as the fluid mechanics can be internal fuel reforming and heat microscale-effective conductivity
modeled simply. In this case, the fuel transfer to the shell-side air. Because could then be used as an input for the
reforming chemistry and charge- the outer tubes acted as radiation full one-dimensional tube-side
transfer electrochemistry inside the shields, the inner tubes generally simulation to further improve transport
tubes are all complex, but the fuel operated at higher temperatures. modeling inside the tubes, thus
flow is indeed simple enough and can Shell-side air was introduced from enabling the beginnings of a
be modeled in a one-dimensional tool below and exhausted at the top, and, multiscale analysis.
as a plug flow. therefore, the oxygen decreased from The capability to couple complex
The next aspect of the overall bottom to top. flow and heat transfer using flow
simulation process was coordinating When designing and optimizing simulation with complex chemistry
the iterative coupling of the cathode- a tubular SOFC stack, and electrochemistry is a powerful
side flow simulation model with simulation is new tool for certain classes of
the anode-side chemistry extraordinarily reacting-flow problems. The CSM
model. Using the ANSYS valuable team developed the ANSYS FLUENT
F LUE N T u s e r- d e f i n e d UDFs to be sufficiently general so that
a range of one-dimensional chemistry
tools could be incorporated. This
a p p ro a c h t o m o d e l i n g t u b u l a r
configurations can be useful beyond
fuel cells, as it is directly applicable to
geometrically related layouts, such as
battery packs, nuclear fuel rods or
A three-dimensional ANSYS FLUENT face mesh on an SOFC tube with an overlying cracking furnaces. n
one-dimensional band mesh

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 47


Tips and Tricks

Accelerating
CFD Solutions
Several recent enhancements in ANSYS FLUENT solver capabilities
accelerate convergence and reduce solution time.
By Mark Keating, Principal Engineer, ANSYS, Inc.

Many solver performance enhancements have been Pseudo-Transient Method


introduced to ANSYS FLUENT fluid dynamics software The pseudo-transient solution method, introduced in
over the past few releases. These capabilities can version 13.0, is a form of implicit under-relaxation for
dramatically improve the speed and reliability of simulation. steady-state cases. It allows users to obtain solutions
Running the solver out of the box does not always faster and more robustly than previous versions of ANSYS
guarantee optimum solver settings for any particular FLUENT software, especially for highly anisotropic
application. So by understanding and using solver tech- meshes, when using PBCS and density-based (DBNS)
nology appropriately, a user can obtain faster results and implicit solvers. This method uses a pseudo-transient
better convergence. time-stepping approach. In general, the time per iteration
is slightly higher, but in some extreme cases the number
Pressure-Based Coupled Solver of iterations required for convergence using this method
The pressure-based coupled solver (PBCS) was has dropped by an order of magnitude or more. Usually,
introduced in 2006, and its usage is growing. This solver overall speedups of 30 percent to 50 percent can be
reduces the time to overall convergence, by as much as expected. Cases with multiple reference frame (MRF)
five times, by solving momentum and pressure-based zones should benefit from using this method. The table
continuity equations in a coupled manner. Though there is shows the levels of improvement possible.
a slight increase in associated memory requirements for
using this solver, its benefits far outweigh the drawbacks. Courant number-
based coupled Pseudo-transient
The PBCS is becoming the solver of choice for subsonic Cases (iterations) coupled (iterations)
applications. When using it, re-ordering the grid is always
Backward facing step 750 75
advisable. The default explicit under-relaxation factors (turbulent: SST)
( UR F s ) f o r p re s s u re a n d v e l o c i t y o f 0 . 7 5 a re Film cooling benchmark 2,300 1,350
generally robust values, but they should be reduced for (turbulent: SA)
skewed meshes, when oscillatory convergence is Flat plate, SST transition 1,200 100
experienced, or when higher-order discretization is model

employed to about 0.4 to 0.5. Taking the turbulence URF Rotor/stator with mixing 500 250
plane model
up to 0.95 to 0.99 can help to accelerate viscous cases.
Centrifugal pump 220 50
To access the PBCS, change the p–v coupling in the
drop-down list from SIMPLE to Coupled. Axial compressor stage 400 110

Solver speedups achieved using the pseudo-transient coupled solver


in ANSYS FLUENT 13.0

48 ANSYS Advantage • © 2011 Ansys, Inc. www.ansys.com


Tips and Tricks

1e+02 1e+02

Residuals
continuity
Residuals x-velocity
continuity 1e+00 y-velocity
1e+00 z-velocity
x-velocity
y-velocity energy
z-velocity
energy 1e-02
1e-02

1e-04

1e-04

1e-06

1e-06
1e-08

1e-08
1e-10

1e-10 1e-12
0 50 100 150 200 250 300 0 20 40 60 80 100 120

Standard Initialization, 279 Iterations Hybrid Initialization, 102 Iterations

Residual plots for a heat exchanger case show the effect on iteration count of hybrid initialization. This case does not converge with
default under-relaxation and standard initialization, but it converges well with the default under-relaxation factors. It is more than
two times faster when hybrid initialization is used.

Initialization Methods solution. It is particularly suited to turbomachinery as well


Providing an initial data field that is close to the final as external and compressible flow problems.
solution for steady-state cases means the solver has to Hybrid solution initialization was introduced at
do less work to reach the converged result. Therefore, this version 13.0 and uses a collection of recipes and boundary
reduces simulation time. Typically, many users employ interpolation methods to efficiently initialize the solution
standard initialization; some use patching for localized based purely on simulation setup — so the user does not
control, especially for moving domains or multiphase need to provide additional inputs for initialization. The
analyses. Interpolation files are used to initialize cases, method can be applied to flows ranging from subsonic to
and the interpolation workflows available in ANSYS supersonic. It is the recommended method when using
FLUENT have been improved in recent releases. There are PBCS and DBNS for steady-state cases in ANSYS FLUENT
also other initialization techniques for further accelerating 13.0. This initialization may improve the convergence
the simulation convergence. robustness for many cases. Unlike FMG, this initialization
Full multigrid initialization (FMG), introduced in 2006, method can be used for multiphase flows.
provides the initial and approximate solution at a minimum
cost to overall computational expense. The feature is Summary
accessed via the text user interface (TUI) and re-orders the A number of solver settings are available to aid
grid as part of the process. The commands are: solution acceleration and convergence within the ANSYS
Solve>initialize>set-fmg-initialization FLUENT solver. Individually, these techniques can be
Solve>initialize>fmg-i used to reduce solution times; combined, they offer even
greater capability. For example, in some cases, moving
The overall initialization time using this approach is from a segregated solver to PBCS with FMG and
much longer than that using standard initialization by the pseudo-transient method has resulted in up
zone, but it allows a much quicker solve. FMG solves to 100 times speedup. These features show the benefits
Euler equations and is available for single-phase flows of investigating and taking advantage of new
only. FMG initialization provides the best-guess initial solver technologies. n

www.ansys.com ANSYS Advantage • Volume V, Issue 1, 2011 49


ANSYS, Inc.
Southpointe
275 Technology Drive
Canonsburg, PA U.S.A. 15317

Send address corrections to


AdvantageAddressChange@ansys.com.